Skip to main content
PMC home page
eLife logoLink to eLife
. 2021 May 7;10:e64711. doi: 10.7554/eLife.64711

Persistent effects of pair bonding in lung cancer cell growth in monogamous Peromyscus californicus

Asieh Naderi 1, Elham Soltanmaohammadi 1, Vimala Kaza 2, Shayne Barlow 3, Ioulia Chatzistamou 4, Hippokratis Kiaris 1,2,✉
Editors: Ivan Topisirovic5, Eduardo Franco6
PMCID: PMC8104960  PMID: 33960931

Abstract

Epidemiological evidence suggests that social interactions and especially bonding between couples influence tumorigenesis, yet whether this is due to lifestyle changes, homogamy (likelihood of individuals to marry people of similar health), or directly associated with host-induced effects in tumors remains debatable. In the present study, we explored if tumorigenesis is associated with the bonding experience in monogamous rodents at which disruption of pair bonds is linked to anxiety and stress. Comparison of lung cancer cell spheroids that formed in the presence of sera from bonded and bond-disrupted deer mice showed that in monogamous Peromyscus polionotus and Peromyscus californicus, but not in polygamous Peromyscus maniculatus, the disruption of pair bonds altered the size and morphology of spheroids in a manner that is consistent with the acquisition of increased oncogenic potential. In vivo, consecutive transplantation of human lung cancer cells between P. californicus, differing in bonding experiences (n = 9 for bonded and n = 7 for bond-disrupted), and nude mice showed that bonding suppressed tumorigenicity in nude mice (p<0.05), suggesting that the protective effects of pair bonds persisted even after bonding ceased. Unsupervised hierarchical clustering indicated that the transcriptomes of lung cancer cells clustered according to the serum donors’ bonding history while differential gene expression analysis pointed to changes in cell adhesion and migration. The results highlight the pro-oncogenic effects of pair-bond disruption, point to the acquisition of expression signatures in cancer cells that are relevant to the bonding experiences of serum donors, and question the ability of conventional mouse models to capture the whole spectrum of the impact of the host in tumorigenesis.

Research organism: Other

eLife digest

People’s social interactions could influence their risk of developing various diseases, including cancer, according to population-level studies. In particular, studies have identified a so-called widowhood effect where a person’s risk of disease increases following the loss of a spouse. However, the cause of the widowhood effect remains debatable, as it can be difficult to separate the impact of lifestyle changes from biological changes in the individual following bereavement.

It is not possible to use laboratory mice to identify a causal biological mechanism, because they do not form long-term relationships with a single partner (pair bonds). However, several species of deer mouse form pair bonds, and suffer from anxiety and stress if these bonds are broken. Naderi et al. used these mice to study the widowhood effect on the risk of developing cancer.

First, Naderi et al. grew human lung cancer cells in blood serum taken from mice that were either in a pair bond or had been separated from their partner. The cancer cells grown in the blood of mice with disrupted pair bonds changed size and shape, indicating that these mice were more likely to develop cancer. This effect was not observed when the cells were grown in the blood of bonded deer mice or of another deer mouse species that does not form pair bonds. Naderi et al. also found that the activity of genes involved in the cancer cells’ ability to spread and to stick together was different in pair-bonded mice and in pair-separated mice.

Next, Naderi et al. implanted lung cancer cells into the deer mice to study their effects on live animals. When cancer cells from the deer mice were transplanted into laboratory mice with a weakened immune system, the cells taken from pair-bonded deer mice were less likely to grow than the cells from deer mice with disrupted pair bonds. This suggests that the protective effects of pair bonding persist even after removal from the original mouse.

These results provide evidence for a biological mechanism of the widowhood effect, where social experiences can alter gene activity relating to cancer growth. In the future, it will be important to determine whether the same applies to humans, and to find out if there are ways to mimic the effects of long-term bonds to improve cancer prognoses.

Introduction

While the psychosomatic impact of cancer in patients is extensively documented, the reciprocal effects of individuals’ social experiences in carcinogenesis receive limited attention. Both anecdotal and experiential evidence, and numerous epidemiological studies, strongly suggest that emotional factors can affect the development and progression of cancer, pointing to the sensitivity of cancer cells to signals associated with behavior, emotional state, and sociality. For example, the marital status modulates the likelihood for the development of fatal cancers, with unmarried, divorced, or widowed individuals exhibiting an increased chance of developing life-threatening disease and males being more susceptible than females to the protective effects of marriage (Aizer et al., 2013). The ‘widowhood effect’ provides an example at which in couples, after the loss of one partner, the surviving one exhibits an increased probability for the development of various fatal pathologies (Elwert and Christakis, 2008; Blanner et al., 2020; Sullivan and Fenelon, 2014; Bowling, 1987; Boyle et al., 2011). Notwithstanding that high variation in death causes has been documented, cancer is recognized as a common cause of mortality (Aizer et al., 2013; Elwert and Christakis, 2008; Blanner et al., 2020; Burgoa, 1998; Martikainen and Valkonen, 1996; Sex, 1973). Although both sexes are influenced by widowhood, males appear more sensitive than females to widowhood-associated death (Sullivan and Fenelon, 2014; Helsing et al., 1981).

Despite the information they provide, unavoidable changes in lifestyle habits in the bereaved partner at widowhood or between single and married patients complicate the epidemiological data interpretation. Several mechanisms connecting cancer to social interactions, mental state, and bereavement have been proposed. Laboratory mice of the genus Mus, despite their power in illuminating various aspects of tumorigenesis, remain of limited value in modeling the effects of pair bonding. It is estimated that in less than 10% of mammals, including humans, individuals form pair bonds that are based on mating (Kleiman, 1977; Lukas and Clutton-Brock, 2013; Scribner et al., 2020). Therefore, mice, by not developing long-term pair bonds, are not adequate in studying the effects of widowhood and pair-bond disruption (Chatzistamou et al., 2018; McDonald et al., 2005). Earlier studies in mice have shown that brain-derived signals linked to the reward system may impact tumorigenesis, whereas stress can stimulate metastases (Ben-Shaanan et al., 2018; Sloan et al., 2010). However, more complex behavioral traits involving social interactions in married couples or widowhood cannot be studied in mice. Peromyscus californicus is a monogamous species developing long-term, cohesive pair bonds that can influence various physiological responses (Havighorst et al., 2017; Perea-Rodriguez et al., 2015; Glasper and Devries, 2005; Wright et al., 2018). Upon cyclosporine-mediated immunosuppression, similarly with other rodents, P. californicus supports the growth of human cancers, providing a potentially informative animal model for the study of pair-bond disruption in tumorigenesis in vivo (Fingert et al., 1984; Kaza et al., 2018; Chatzistamou and Kiaris, 2016).

Results

Bonding history modulates the effects of sera in tumor spheroid formation

Initially, we asked if sera of P. californicus following the disruption of pair bonds affected the growth of cancer cells in vitro in a manner that depended on bonding history. We focused on the formation of tumor spheroids that are enriched in cells with cancer stem cell (CSC)-like properties, and their formation is known to reflect tumorigenic activity directly (Visvader and Lindeman, 2012; Ishiguro et al., 2017). Sera were obtained from 14 to 17 months old virgin, bonded for about 12 months, or bond-disrupted (after 12 months of bonding) at the periods indicated, male P. californicus, and the efficacy of spheroid formation by A549 human lung cancer cells was assessed. A pilot study indicated that sera obtained from animals 1 week after the disruption of pair bonds resulted in the formation of larger yet less compact spheroids, suggesting a significant impact of bond disruption in spheroid morphogenesis (Figure 1a). The results were confirmed and extended in a subsequent study that also included sera obtained 24 hr and 2 weeks after the disruption of pair bonds (Figure 1b). In this study, sera from 9 (B), 5 (BD, 24 hr), 5 (BD, 1 week), 4 (BD, 2 weeks), and 5 (virgin, V) different animals were used, and microsphere formation was evaluated in two biological replicas for each (n = 10 for BD [1 week], BD [24 hr], and V; n = 8 for BD [2 weeks] and n = 18 for B). For control media (CM) and plain serum-free media (PM), n = 4. As shown in Figure 1b, this activity was only marginal at 24 hr but was significant (p<0.05) 1 week and 2 weeks after the disruption of pair bonds, implying that the factors responsible accumulated in the sera after pair-bond disruption. As compared to virgins, sera from animals at bonding resulted in the formation of smaller spheroids, albeit insignificantly, which implies that bonding may also have some protective activity, beyond the pro-oncogenic activity of bond disruption (Figure 1b).

Figure 1. Effects of pair bonding in the pro-oncogenic activity of sera.

(a) Representative microphotographs of tumor spheroids developed by A549 cells 3 days after cell seeding. Cells formed spheroids in the presence of sera from bonded (B), bond-disrupted animals (BD) 1 week after disruption, and virgin animals (V), and control media containing fetal bovine serum (FBS) (CM). Live (green) and dead cells (red) are indicated. Representative images of two independent experiments are shown. (b) Representative microphotographs of tumor spheroids developed by A549 cells on day 1, day 2, and day 3, after cell seeding. Cells formed spheroids in the presence of sera from bonded (B), bond-disrupted animals (BD) 24 hr, 1 week, and 2 weeks after disruption, and virgin animals (V), control media containing FBS (CM) or serum-free plain media (PM). The last column shows images at day three in higher magnification. Bars indicate 200 μM. (c) Scatter dot plots of data shown in (b), indicating the size of tumor spheroids at days 2 and 3 after seeding. Median and p-values are indicated. Sera from 9 (B), 5 (BD, 24 hr), 5 (BD, 1 week), 4 (BD, 2 weeks), and 5 (V) different animals were used and microsphere formation was evaluated in two biological replicas for each (n = 10 for BD [1 week], BD [24 hr] and V; n = 8 for BD [2 weeks] and n = 18 for B). For control media (CM) and plain serum-free media (PM), n = 4. For this experiment, sera from 14 to 17 months old mice that were used for the B and BD groups were bonded for about 1 year. Statistical analyses were performed by ANOVA.

Figure 1.

Figure 1—figure supplement 1. Microsphere morphology of a panel of human lung cancer cells cultured in P. californicus sera.

Figure 1—figure supplement 1.

Open in a new tab

Variation in spheroid size with sera obtained after bond disruption is due to the genetic diversity of donor animals and persists in different lung cancer cell lines

 The effects in spheroid size described above were obtained with sera from older animals (14–17 months old) that were bonded for at least 12 months. To test whether disruption of bonds in younger animals that were bonded for shorter time periods also produced similar effects, we conducted the following study: We exposed to sera of 8–10 months old animals that were either pair-bonded for 2 months or following 2 weeks of bond disruption after 2 months of bonding, a roster of lung cancer cell lines. For this experiment, 14 animals were used that represented seven sibling pairs with each sibling allocated either to the bonded or to the bond-disrupted group. Our results indicated that consistently, in the same sibling pair, an induction of microsphere size of similar magnitude was noted for all five cell lines tested, in four of seven pairs, while this effect was only marginal in the remaining three pairs (Figure 1—figure supplement 1). The variation in spheroid size was analogous to that recorded in the results described in Figure 1c. Thus, we conclude that even shorter periods of bonding are sufficient, and the consequences of its disruption can be recorded in sera from even younger animals. More importantly, it indicates that the variation of the effects is due to the diversity of the animals and not to the differential sensitivity of the cell lines used.

Persistent pro-oncogenic activity of bond disruption in vivo

The effects of pair bonding in spheroid formation prompted us to explore whether bond disruption also influences the efficacy of tumorigenesis in vivo. To that end, vasectomized male P. californicus were allowed to establish pair bonds for about 2 months with their female partners and then subjected to pair-bond disruption (n = 9) or were left with their partners (n = 11). Following immunosuppression by CsA animals were inoculated with A549 human lung cancer cells and tumorigenesis was monitored. Animals that did not possess bonding experiences before were used as controls (n = 8). Tumors grew originally in animals of all experimental groups and by day 15 measurable tumors were detected in 9 of 11 bonded, in 8 of 9 bond-disrupted and in 6 of 8 virgins (Figure 2a). At this point, tumors were modestly – albeit not statistically significantly – larger in the bond-disrupted animals and smaller in the group of virgins (Figure 2a). By day 25, the tumors persisted in both the bond-disrupted and bonded animals, at 89% (8 of 9) and 82% (9 of 11) rate, respectively, while in virgin animals, they were detectable only in 25% (2 of 8) of the animals (Figure 2b,c).

Figure 2. Growth of A549 human lung cancers in P. californicus (IS stock) and bonding experience.

Figure 2.

Open in a new tab

Vasectomized males were used in all studies. (a) Volume of measurable tumors at day 15 following cancer cell inoculation in the bonded (n = 9), bond-disrupted (n = 8), and virgin (n = 6) groups; out of the 11, 9 and 8 animals implanted originally with A549 cells. (b) Pie graphs indicating the percentage of animals bearing tumors at day 25. (c) Representative hematoxylin and eosin (H and E)-stained sections of P. californicus-grown tumors from bonded (upper panel), and bond-disrupted (lower panel) groups. (*) and (#) indicate necrotic areas and muscle invasion, respectively. (d) Tumor-free nude mice implanted with A549 tumor explants from bonded (n = 9) and bond-disrupted (n = 7) P. californicus. p-values (log-rank [Mantel–Cox] test) are shown. (e) H and E-stained sections of A549 tumors in nude mice derived from explants of A549 tumors from bonded, and bond-disrupted P. californicus. The morphology of A549 tumors from the direct inoculation of A549 cells in nude mice is shown (n). N is indicated in the text. B-IS, tumors growing in bonded P. californicus; BD-IS, tumors growing in bond-disrupted P. californicus; B-n, tumors that originally developed in bonded P. californicus and now growing in nude mice; BD-n, tumors that originally developed in bond-disrupted P. californicus and now growing in nude mice; n, tumors that developed in nude mice following injection of A549 cells.

In a follow-up study, we explored if differential pro-oncogenic activity persisted after growth in nude mice. Thus, tumors that were originally grown in P. californicus for at least 1 month (n = 9 for bonded and n = 7 for bond-disrupted) were re-transplanted in virgin nude mice (one nude mice for each original Peromyscus tumor), and tumorigenesis was recorded. As shown in Figure 2d, tumors from bonded P. californicus exhibited significantly (p=0.011) lower tumorigenicity in nude mice than those grown originally in the bond-disrupted animals, despite that histologically they remained indistinguishable (Figure 2e). In line with the tumor spheroid analyses, pair bonding produced persistent changes in tumors that suppressed their growth and endured even when bonding seized.

Effects of pair bonding in differential gene expression

The effects of bonding history in the profile of tumor growth in vivo, combined with the spheroid formation in vitro, imply the induction of transcriptional changes in the cancer cells in a manner that depends on bonding experience (Figure 3, Figure 3—figure supplements 1, 2 and 3). Initially, we focused on the expression of established CSC markers and genes regulating CSC potential, such as Oct-4, b-catenin, and CD-133 that have been identified previously in A549 cells (Chiou et al., 2010; Akunuru et al., 2012; Teng et al., 2010). The analysis was performed by semiquantitative RT-PCR in 2D cultures to eliminate the effects of the clonal selection of cells in the spheroids. Differential expression analysis did not reveal considerable differences between the bonding groups, either in cells cultured in vitro with sera from animals differing in bonding history or in vivo in tumors in nude mice or Peromyscus (Figure 3—figure supplement 1). However, unsupervised hierarchical clustering indicated that these CSC markers provided a signature that predicted a relatively high accuracy bonding history of the animals (Figure 3—figure supplement 1).

Figure 3. RNAseq analysis of A549 cells cultured in the presence of sera from bonded (B), bond-disrupted (BD), or virgin (V) P. californicus.

(a) Bar graphs showing number of differentially expressed genes in each pairwise comparison group. (b ) Volcano plots showing differentially expressed genes between the B vs BD, and B vs V groups. (c) Venn Diagrams showing overlapping differentially expressed genes. The identity of genes is shown in the right. In the bonded group, mice were paired for 12 months. For the bond-disrupted group, we separated paired mice after 12 months of bonding, and collected the sera 1 week after bond disruption. For virgin mice, we collected sera from mice housed 3/cage.

Figure 3.

Figure 3—figure supplement 1. CSC markers in A549 cells cultured in vitro and tumors.

Figure 3—figure supplement 1.

Figure 3—figure supplement 2. Unsupervised hierarchical clustering based on log2 transformed values on all RNAseq data.

Figure 3—figure supplement 2.

Figure 3—figure supplement 3. RNAseq analysis summary of A549 cells.

Figure 3—figure supplement 3.

Open in a new tab

This observation prompted us to perform RNA sequencing and analyze expression profiles at the whole transcriptome level in human A549 lung cancer cells in the presence of sera that had been isolated from monogamous male P. californicus that were virgin (V), bonded (B), or subjected to disruption of pair bonds (BD) after bonding (n = 6 samples/group). Controls (C) cultured in the presence of fetal bovine serum (FBS) were also included. Unsupervised hierarchical clustering (Vidman et al., 2019) indicated that the transcriptomes clustered well together according to the serum donors’ bonding history, except the virgin (V) group that exhibited the lowest discrimination (Figure 3—figure supplement 2). Differential gene expression analysis was performed as described before by using the iDEP platform (Ge et al., 2018). This analysis showed that the majority of differentially expressed genes were detected in the comparisons involving the FBS-treated cells (C), which suggests that the species origin of sera produces the most potent effects in gene expression and potentially masking the consequences of pair bonding in the regulation of the transcriptome (Figure 3—figure supplement 3). Thus, we repeated the analysis by excluding the specimens corresponding to FBS and restricted it only to the specimens that received Peromyscus sera (Figure 3). Seven genes were differentially expressed in each B vs BD and B vs V comparisons, while none were detected between the V and BD groups (Table 1). Thus, it seems that pair bonding produces more robust effects in the sera as compared to those of bond disruption. Among these genes, all of which were downregulated in the B group, five were common and included HES1, ZFP36, NR4A1, FGG, and SOCS3. Hes1 is a transcription factor that is downstream of Notch signaling, for which the pro-oncogenic activity in lung cancer has been established (Westhoff et al., 2009; Yuan et al., 2015). NR4A1 encodes for the orphan nuclear receptor A1 for which a strong association with unfavorable outcome in lung cancer has been shown and is involved in cancer cell migration (Zhu et al., 2017; Hedrick et al., 2018). SOCS3 is a suppressor of cytokine signaling and is a repressor of lung tumorigenesis (He et al., 2003; Lund and Rigby, 2006). FGG encodes for fibrinogen gamma chain that has been linked to enhanced invasion of lung and other cancer cells (Sahni et al., 2008; Zhang et al., 2019). The genes that were uniquely detected in the BD vs B groups comparison were FGA and FGB, which encode for fibrinogen A and B chains (Pieters and Wolberg, 2019), while in the V vs B comparison, the oncogene Jun that enhances lung cancer cell migration (Shimizu et al., 2008) and the connective tissue growth factor that at least in lung cancer, is associated with favorable prognosis (Chien et al., 2006; Chang et al., 2004). Pathway enrichment analysis indicated that processes associated with differentially expressed genes were linked to the regulation of cell migration and spread, or tissue morphogenesis (Table 2).

Table 1. Differentially expressed genes between the B, BD, and V groups.

Chromosomal location, fold change (log2), and adjusted p-value are indicated. Genes that are common in the B vs BD and V vs B comparisons are underlined.

Symbol Chr log2 fold change Adj.pval
BD vs B
 FGG 4q32.1 1.49 8.82e-04
 FGA 4q31.3 1.37 8.82e-04
 FGB 4q31.3 1.24 2.78e-03
 HES1 3q29 1.16 3.36e-06
 NR4A1 12q13.13 1.13 1.42e-09
 SOCS3 17q25.3 1.09 4.39e-10
 ZFP36 19q13.2 1.02 1.77e-03
V vs B
 CTGF 6q23.2 1.26 5.68e-02
 HES1 3q29 1.18 2.26e-06
 ZFP36 19q13.2 1.16 9.80e-05
 NR4A1 12q13.13 1.12 3.41e-09
 FGG 4q32.1 1.08 5.39e-02
 JUN 1p32.1 1.04 3.36e-02
 SOCS3 17q25.3 1.04 3.54e-09
Open in a new tab

Table 2. Biological processes associated with the differentially expressed genes in B vs BD and V vs B groups (He et al., 2003).

The adjusted p-values are indicated.

Group comparison Adj.pval Biological process
BD vs B 3.1e-06 Positive regulation of substrate adhesion-dependent cell spreading
V vs B 4.9e-04 Blood vessel development
5.3e-04 Positive regulation of intracellular signal transduction
5.3e-04 Anatomical structure morphogenesis
5.3e-04 Regulation of cellular protein metabolic process
5.3e-04 Negative regulation of apoptotic process
5.3e-04 Positive regulation of cell differentiation
5.3e-04 Regulation of epithelial cell proliferation
Open in a new tab

Monogamous and not polygamous Peromyscus are sensitive to the effects of bond disruption in spheroid formation

The findings on P. californicus prompted us to explore whether other Peromyscus species are also sensitive to the effects of the disruption of pair bonds. Thus, we compared the effects of sera from bonded or bond-disrupted polygamous P. maniculatus and monogamous Peromyscus polionotus, in the size and shape of A549 tumor spheroids. As shown in Figure 4, the disruption of pair bonds altered spheroid morphology in the monogamous, but not in the polygamous species. The intensity of this effect was variable among the animals tested and was recorded in at least 6 of 12 male P. polionotus but none of P. maniculatus (n = 12) tested (p=0.005, chi-square test; Figure 4—figure supplement 1). Contrary to P. californicus though, at which pair-bond disruption enhanced spheroid size, in P. polionotus the primary effect was seen in the spheroids’ shape: Spheroids that formed in the presence of P. polionotus sera obtained after the disruption of pair bonds had scattered morphology, as opposed to the spheroids from P. maniculatus sera at bonding and bond disruption and those of P. polionotus at bonding that were smooth-edged. In some instances (about 25% of animals), this scattered phenotype was also noted in P. polionotus sera obtained at bonding (Figure 4—figure supplement 1). Whether this difference represents the actual phenotypic difference between the two species or is due to methodological changes in the state of the cells and donor animals remains to be established. In addition, it may reflect the same effect (cell dispersion followed by proliferation) but recorded at different stages during the formation of the spheroids. It is also noted, that the monogamous behavior in Peromyscus has developed independently during the evolution of P. polionotus and P. californicus, and thus alternative signaling ques may have been engaged in altering the consequences of bond disruption in spheroid formation (Jašarević et al., 2013). To that end, the signaling cascades influencing spheroid size and shape may be distinct for the two species; nevertheless, the effects of pair-bond disruption persist.

Figure 4. Tumor spheroids with sera from P. polionotus and P. maniculatus.

Morphology of tumor spheroids formed after 2, 3, and 7 days in culture with sera isolated from monogamous P. polionotus and polygamous P. maniculatus. In 6 of 12 P. polionotus but none of 12 P. maniculatus enhanced dispersion was noted at bond disruption. For the experiment, sera were obtained from the same animal at bonding for 12 months and after 1 week following bond disruption. Bond-disrupted animals were housed independently after separation from females. BW B, P. maniculatus bonded; BW BD, P. maniculatus bond-disrupted; PO B, P. polionotus bonded; PO BD, P. polionotus bond disrupted.

Figure 4.

Figure 4—figure supplement 1. Microsphere morphology of A549 human lung cancer cells cultured in P. maniculatus or P. polionotus sera.

Figure 4—figure supplement 1.

Open in a new tab

Discussion

The present findings exemplify the role of the context – in its wider sense – in cancer progression and underscore the significance of psychosomatic factors as modulators of cancer growth. Using a behaviorally relevant animal model, our results highlight the biological basis of the ‘widowhood effects’ and suggest that it operates as a tumor-promoting factor, beyond lifestyle changes. Our conclusions are based on the recorded effects of pair bonding in three major phenotypic characteristics of the cancer cells. Those included tumor spheroid formation established in the presence of sera from bond-disrupted animals, the expression profile of the cancer cells in vitro and in vivo that depended on the bonding history of serum donors and tumor hosts, respectively, and ultimately their tumorigenicity in the nude mice. The use of sera from outbred, genetically diverse rodents, allowed us to obtain evidence that this effect varies among individuals but persists across different lung cancer cells. This observation might be of relevance to the study of human populations that are genetically diverse and their responses to the same stimuli may be variable.

In our animal model, cancer cells were implanted in tumor-free animals and the kinetics of tumorigenesis was affected by the animals’ bonding history. Whether pair bonding and disruption can also influence tumor initiation will have to be established, nevertheless, the fact that most cancers are slow-growing in patients is consistent with the effects of widowhood in influencing the progression, as opposed to the initiation of the disease. Yet, by using the in vivo experiments immunocompromised animals (nude mice and cyclosporine administration), our study suffers from the absence of integration of immune responses that may be especially relevant to widowhood-associated stress.

An unexpected finding was the loss of the tumors in the virgin animals as opposed to the majority of the bonded and bond-disrupted that retained them (Figure 2b). A possible explanation is probably related to the differential effectiveness of immunosuppression by cyclosporine. Especially during the initial period after cancer cell inoculation, cyclosporine may have caused more potently immunosuppression in the animals that had been subjected to bonding, due to the concomitant anti-inflammatory action of oxytocin, a neurohormone with essential role in the establishment of social interactions and pair bonding (Lutgendorf et al., 2005; Fagundes et al., 2011; Fuligni et al., 2009; Yuan et al., 2016; Carter and Perkeybile, 2018). It is noted though that the high difference in the tumorigenicity between virgins and the bonded or bond-disrupted animals, renders differential immune suppression unlikely as the sole contributor for this discrepancy.

Differential analysis of gene expression showed that sera from animals at bonding enriched for genes regulating cell migration and spreading, and tissue morphogenesis, features that are consistent with the recorded changes in spheroid morphology. Although for several of the differentially expressed genes, their downregulation, which was seen in the bonded group, was associated with a favorable prognosis, in some cases, it was not. For example, SOCS3 was downregulated in the bonding group, yet it is a tumor suppressor for lung and other cancers (He et al., 2003; Lund and Rigby, 2006), which may reflect responses related to oxytocin signaling during bonding (Matarazzo et al., 2012).

Beyond its effects in the expression of individual genes, the impact of bonding history in transcription was more clearly reflected in the similarity recorded in the transcriptomic profiles of cells cultured in sera from animals with similar bonding experiences. This was especially pertinent to the bonded and bond-disrupted groups. An intriguing possibility is that this is indicative for the lowest rigidity in the transcriptomic profile induced by the serum of virgin animals, as opposed to the changes triggered by the sera of bonded and of bond-disrupted animals that remained more robust.

Collectively, the results provide a mechanistic foundation for the widowhood effect and suggest that the individuals’ social, and especially bonding experiences, modify the transcriptome of lung tumors modulating oncogenic activity. As such, they advocate that cancers at widowhood represent a distinct pathological entity that may deserve targeted therapeutic strategies, which should take into consideration social interactions. Thus, preventive measures could be developed to mitigate such pro-oncogenic effects in individuals at bereavement. Whether these findings do occur and at which extent in other monogamous species, including humans, and whether they are applicable to other cancers as well as other pathologies beyond malignancy, remains to be explored. Finally, the present results also raise some concerns regarding the use of conventional animal models and their ability to accurately capture the whole spectrum of the tumorigenic process and the associated host-derived factors.

Materials and methods

Animal studies

Genetically diverse male P. californicus (stock IS), P. polionotus (PO stock), and P. maniculatus (BW stock) were obtained from the Peromyscus Genetic Stock Center (Columbia, SC) (RRID:SCR_002769). Mice were all 14–17 months old and were divided into three experimental groups: bonded, bond-disrupted, and virgin. For the tumor inoculation studies, in the bonded group, mice were paired for at least 2 months before the study began and remained paired until the end of the study. In the bond-disrupted group, mice were paired 2 months before the study started, and immediately after cancer cells injection, they were separated. In the virgin group, mice were kept individually 2 months before the study began. Vasectomy was performed to prevent pregnancy during the study. Some siblings were used and were distributed randomly in different experimental groups as described in the legend of Figure 1—figure supplement 1. Nude mice (male, 6–8 weeks old) were obtained from Charles River Laboratories (Boston, MA) and were housed in groups of 4–5. For serum collection used in the RNAseq studies and spheroid formation, for the bonded group, mice were paired for about 12 months. For the bond-disrupted group, we separated paired mice after 12 months of bonding and collected the sera 1 week after bond disruption. For virgin mice, we collected sera from mice housed 3/cage. Sera were obtained by retro-orbital bleeding before and after bond disruption at the indicated times. Animal studies were approved by the University of South Carolina IACUC (Protocol # 2473-101464-102319).

Cell lines

A549 human non-small cell lung adenocarcinoma cells were originally obtained from ATCC (Manassas, VA) and thereafter maintained in freezing media (60% Dulbecco’s modified Eagle medium [DMEM], 30% FBS, 10% dimethyl sulfoxide). Most recently, cells were validated by STR typing (Biosynthesis, Lewisville, TX) just after completion of experiments. Human H1703 squamous, H596 adenosquamous, H358 bronchioalveolar, and H292 mucoepidermoid lung cancer cells were obtained prior to their use from ATCC (Manassas, VA) and cultured for three passages at ATCC-recommended media prior to the performance of the spheroid assays. Cells were tested negative for mycoplasma contamination.

Tumor inoculation

To cause immunosuppression in P. californicus and overcome xenograft rejection, animals were treated daily with 100 mg/kg cyclosporine A (in 90% olive oil and 10% EtOH) s.c. starting 1 day before the implantation of cancer cells, for 2 weeks, and then every other day for the whole duration of the study (Perea-Rodriguez et al., 2015). For cancer cell inoculation, (5 × 106) cells were injected subcutaneously into the right flank of mice in a total volume of 100 μl phosphate-buffered solution (PBS). Tumor volumes were assessed by using the following formula: (width)2 × length/2. All experiments were approved by the Institutional Animal Care and Use Committee of the University of South Carolina (approval no. 101464). For re-transplantation in nude mice, tumors were harvested from P. californicus, mechanically minced at pieces of 5–10 mm3, and were implanted into the right flank of nude mice using a trocar needle. Mice were followed up each week until 4 months.

Histology

Tumor was fixed in 4% neutral buffered formalin and subsequently embedded in paraffin. Sections were stained with hematoxylin and eosin for histological assessment. Where available, a part of the tumor was snap-frozen on dry ice and stored at −80°C, for RNA extraction. Images were obtained by a Leica optical microscope.

Tumor spheroid formation

Lung cancer cells were seeded into 96-well spheroid microplates (Corning Cat. No. 4515) at 2 × 103 cells/well in 100 μl of DMEM+5% FBS+5% serum of each mouse. The age of the mice, their bonding group, and the period of bonding are described in the text and corresponding figure legends. The plate was incubated at 37°C, 5% CO2. Images were taken using an inverted microscope at 4× magnification each day until 3 days and analyzed using NIH ImageJ software to assess microsphere areas and volumes. The studies were repeated independently at least twice, and similar results were obtained. For the assessment of the spheroids that formed with P. polionotus sera, ‘scattered’ phenotype was scored when at least two outgrowths formed distal from the main spheroid.

Cell viability assay

Spheroid cell viability was assayed using the LIVE/DEAD Viability/Cytotoxicity Kit (Cat. No. L3224). After 3 days of spheroid culture, wells were rinsed two times with an 80 percent-volume change of media with D-PBS. EthD-1 (12 μM) and calcein AM (4 μM) were added to the wells, and the cells were incubated in the dark for 30–45 min to avoid the photodynamic effect. Images were taken using a fluorescence microscope; live cells fluoresce green, whereas dead cells fluoresce red. Data were analyzed using ImageJ image analysis software.

Quantitative real-time PCR analysis and RNA sequencing

Total RNA from cell and tumor tissues were isolated using the Qiagen RNeasy Mini Kit. Equal quantities of RNA were used for making cDNA using iScript cDNA synthesis kits (Bio-Rad) according to the supplier’s protocol on a T100 thermal cycler (Bio-Rad). Human-specific primers for CSC-related genes: OCT4, β-catenin, and CD133 were designed using Primer3 and Primer BLAST. Quantitative real-time PCR was performed using the Bio-Rad Real-Time PCR detection system and iTaq Universal SYBR Green Supermix (Bio-Rad) according to the manufacturer’s instructions. Amounts of target genes mRNA were normalized to a reference gene GAPDH and were expressed as arbitrary units. The oligonucleotides used for qPCR amplification were as follows: Oct-4: GAAGGATGTGGTCCGAGTGT (left) and GTGAAGTGAGGGCTCCCATA (right); b-catenin: GAAACGGCTTTCAGTTGAGC (left) and CTGGCCATATCCACCAGAGT (right); CD-133: TTGTGGCAAATCACCAGGTA (left) and TCAGATCTGTGAACGCCTTG (right); GAPDH: CCATCACCATCTTCCAGGAGCG (left) and AGAGATGATGACCCTTTTGGC (right). Hierarchical clustering analysis and presentation of expression data were performed using the Morpheus analysis software (https://software.broadinstitute.org/morpheus). For the analysis, raw cpm values were either used or transformed by using the formula Log2 (1 + raw values), as described in the text. RNA sequencing was performed as described (Chavez et al., 2020). RNAseq data were deposited to NCBI (GSE167827). Differential analysis of gene expression and enrichment pathway analysis were performed by using the iDEP platform (He et al., 2003).

Statistical analysis

The data are presented as mean ± SEM. Statistical analysis was performed by paired t-test, chi-square test, ANOVA, or log-rank (Mantel–Cox) test as indicated in the figure legends and text. Results were considered significant when p≤0.05. All graphs were generated using GraphPad Prism software (version 8).

Acknowledgements

The results shown here are in part based upon data generated by the TCGA Research Network: https://www.cancer.gov/tcga. We thank Hao Ji and Dr Michael Shtutman form the UofSC Functional Genomics Core for the RNA sequencing analysis and Dr Vitali Sikirzhytski for help with fluorescent imaging. This study was supported by NSF (Award Number: OIA1736150).

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Contributor Information

Hippokratis Kiaris, Email: kiarish@cop.sc.edu.

Ivan Topisirovic, Jewish General Hospital, Canada.

Eduardo Franco, McGill University, Canada.

Funding Information

This paper was supported by the following grant:

  • National Science Foundation OIA-1736150 to Hippokratis Kiaris.

Additional information

Competing interests

No competing interests declared.

Author contributions

Data curation, Formal analysis, Investigation, Methodology, Writing - original draft, Writing - review and editing.

Data curation, Investigation, Writing - review and editing.

Resources, Investigation, Writing - review and editing.

Investigation, Writing - review and editing.

Conceptualization, Investigation, Writing - review and editing.

Conceptualization, Formal analysis, Supervision, Funding acquisition, Writing - original draft, Project administration, Writing - review and editing.

Ethics

Animal experimentation: University of South Carolina IACUC (Protocol # 2473-101464-102319).

Additional files

Transparent reporting form

Data availability

Peromyscus animals are available from the Peromyscus Genetic Stock Center.

The following dataset was generated:

Naderi A, Ji H, Shtutman M, Kiaris H. 2021. Whole Transcriptome RNA-seq from A549 cells exposed to P. californicus serum. NCBI Gene Expression Omnibus. GSE167827

References

  1. Aizer AA, Chen MH, McCarthy EP, Mendu ML, Koo S, Wilhite TJ, Graham PL, Choueiri TK, Hoffman KE, Martin NE, Hu JC, Nguyen PL. Marital status and survival in patients with Cancer. Journal of Clinical Oncology. 2013;31:3869–3876. doi: 10.1200/JCO.2013.49.6489. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Akunuru S, James Zhai Q, Zheng Y. Non-small cell lung Cancer stem/progenitor cells are enriched in multiple distinct phenotypic subpopulations and exhibit plasticity. Cell Death & Disease. 2012;3:e352. doi: 10.1038/cddis.2012.93. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Ben-Shaanan TL, Schiller M, Azulay-Debby H, Korin B, Boshnak N, Koren T, Krot M, Shakya J, Rahat MA, Hakim F, Rolls A. Modulation of anti-tumor immunity by the brain's reward system. Nature Communications. 2018;9:2723. doi: 10.1038/s41467-018-05283-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Blanner C, Mejldal A, Prina AM, Munk-Jørgensen P, Ersbøll AK, Andersen K. Widowhood and mortality: a danish nationwide register-based cohort study. Epidemiology and Psychiatric Sciences. 2020;29:E149. doi: 10.1017/S2045796020000591. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bowling A. Mortality after bereavement: a review of the literature on survival periods and factors affecting survival. Social Science & Medicine. 1987;24:117–124. doi: 10.1016/0277-9536(87)90244-9. [DOI] [PubMed] [Google Scholar]
  6. Boyle PJ, Feng Z, Raab GM. Does widowhood increase mortality risk? Epidemiology. 2011;22:1–5. doi: 10.1097/EDE.0b013e3181fdcc0b. [DOI] [PubMed] [Google Scholar]
  7. Burgoa M. Mortality by cause of death and marital status in Spain. The European Journal of Public Health. 1998;8:37–42. doi: 10.1093/eurpub/8.1.37. [DOI] [Google Scholar]
  8. Carter CS, Perkeybile AM. The monogamy paradox: what do love and sex have to do with it? Frontiers in Ecology and Evolution. 2018;6:202. doi: 10.3389/fevo.2018.00202. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Chang C-C, Shih J-Y, Jeng Y-M, Su J-L, Lin B-Z, Chen S-T, Chau Y-P, Yang P-C, Kuo M-L. Connective tissue growth factor and its role in lung adenocarcinoma invasion and metastasis. JNCI: Journal of the National Cancer Institute. 2004;96:364–375. doi: 10.1093/jnci/djh059. [DOI] [PubMed] [Google Scholar]
  10. Chatzistamou I, Farmaki E, Kaza V, Kiaris H. The value of outbred rodent models in Cancer research. Trends in Cancer. 2018;4:468–471. doi: 10.1016/j.trecan.2018.05.004. [DOI] [PubMed] [Google Scholar]
  11. Chatzistamou I, Kiaris H. Modeling estrogen receptor-positive breast cancers in mice: is it the best we can do? Endocrine-Related Cancer. 2016;23:C9–C12. doi: 10.1530/ERC-16-0397. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Chavez B, Farmaki E, Zhang Y, Altomare D, Hao J, Soltnamohammadi E, Shtutman M, Chatzistamou I, Kiaris H. A strategy for the identification of paracrine regulators of Cancer cell migration. Clinical and Experimental Pharmacology & Physiology. 2020;47:1758–1763. doi: 10.1111/1440-1681.13366. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Chien W, Yin D, Gui D, Mori A, Frank JM, Said J, Kusuanco D, Marchevsky A, McKenna R, Koeffler HP. Suppression of cell proliferation and signaling transduction by connective tissue growth factor in non-small cell lung Cancer cells. Molecular Cancer Research. 2006;4:591–598. doi: 10.1158/1541-7786.MCR-06-0029. [DOI] [PubMed] [Google Scholar]
  14. Chiou SH, Wang ML, Chou YT, Chen CJ, Hong CF, Hsieh WJ, Chang HT, Chen YS, Lin TW, Hsu HS, Wu CW. Coexpression of Oct4 and nanog enhances malignancy in lung adenocarcinoma by inducing Cancer stem cell-like properties and epithelial-mesenchymal transdifferentiation. Cancer Research. 2010;70:10433–10444. doi: 10.1158/0008-5472.CAN-10-2638. [DOI] [PubMed] [Google Scholar]
  15. Elwert F, Christakis NA. The effect of widowhood on mortality by the causes of death of both spouses. American Journal of Public Health. 2008;98:2092–2098. doi: 10.2105/AJPH.2007.114348. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Fagundes CP, Bennett JM, Derry HM, Kiecolt-Glaser JK. Relationships and inflammation across the lifespan: social developmental pathways to disease. Social and Personality Psychology Compass. 2011;5:891–903. doi: 10.1111/j.1751-9004.2011.00392.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Fingert HJ, Treiman A, Pardee AB. Transplantation of human or rodent tumors into cyclosporine-treated mice: a feasible model for studies of tumor biology and chemotherapy. PNAS. 1984;81:7927–7931. doi: 10.1073/pnas.81.24.7927. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Fuligni AJ, Telzer EH, Bower J, Cole SW, Kiang L, Irwin MR. A preliminary study of daily interpersonal stress and C-reactive protein levels among adolescents from latin american and european backgrounds. Psychosomatic Medicine. 2009;71:329–333. doi: 10.1097/PSY.0b013e3181921b1f. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Ge SX, Son EW, Yao R. iDEP: an integrated web application for differential expression and pathway analysis of RNA-Seq data. BMC Bioinformatics. 2018;19:534. doi: 10.1186/s12859-018-2486-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Glasper ER, Devries AC. Social structure influences effects of pair-housing on wound healing. Brain, Behavior, and Immunity. 2005;19:61–68. doi: 10.1016/j.bbi.2004.03.002. [DOI] [PubMed] [Google Scholar]
  21. Havighorst A, Crossland J, Kiaris H. Peromyscus as a model of human disease. Seminars in Cell & Developmental Biology. 2017;61:150–155. doi: 10.1016/j.semcdb.2016.06.020. [DOI] [PubMed] [Google Scholar]
  22. He B, You L, Uematsu K, Zang K, Xu Z, Lee AY, Costello JF, McCormick F, Jablons DM. SOCS-3 is frequently silenced by hypermethylation and suppresses cell growth in human lung Cancer. PNAS. 2003;100:14133–14138. doi: 10.1073/pnas.2232790100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hedrick E, Mohankumar K, Safe S. TGFβ-Induced lung Cancer cell migration is NR4A1-Dependent. Molecular Cancer Research. 2018;16:1991–2002. doi: 10.1158/1541-7786.MCR-18-0366. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Helsing KJ, Szklo M, Comstock GW. Factors associated with mortality after widowhood. American Journal of Public Health. 1981;71:802–809. doi: 10.2105/AJPH.71.8.802. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Ishiguro T, Ohata H, Sato A, Yamawaki K, Enomoto T, Okamoto K. Tumor-derived spheroids: relevance to Cancer stem cells and clinical applications. Cancer Science. 2017;108:283–289. doi: 10.1111/cas.13155. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Jašarević E, Bailey DH, Crossland JP, Dawson WD, Szalai G, Ellersieck MR, Rosenfeld CS, Geary DC. Evolution of monogamy, paternal investment, and female life history in Peromyscus. Journal of Comparative Psychology. 2013;127:91–102. doi: 10.1037/a0027936. [DOI] [PubMed] [Google Scholar]
  27. Kaza V, Farmaki E, Havighorst A, Crossland J, Chatzistamou I, Kiaris H. Growth of human breast cancers in Peromyscus. Disease Models & Mechanisms. 2018;11:dmm031302. doi: 10.1242/dmm.031302. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Kleiman DG. Monogamy in mammals. The Quarterly Review of Biology. 1977;52:39–69. doi: 10.1086/409721. [DOI] [PubMed] [Google Scholar]
  29. Lukas D, Clutton-Brock TH. The evolution of social monogamy in mammals. Science. 2013;341:526–530. doi: 10.1126/science.1238677. [DOI] [PubMed] [Google Scholar]
  30. Lund PK, Rigby RJ. SOC-ing it to tumors: suppressors of cytokine signaling as tumor repressors. Gastroenterology. 2006;131:317–319. doi: 10.1053/j.gastro.2006.05.030. [DOI] [PubMed] [Google Scholar]
  31. Lutgendorf SK, Sood AK, Anderson B, McGinn S, Maiseri H, Dao M, Sorosky JI, De Geest K, Ritchie J, Lubaroff DM. Social support, psychological distress, and natural killer cell activity in ovarian Cancer. Journal of Clinical Oncology. 2005;23:7105–7113. doi: 10.1200/JCO.2005.10.015. [DOI] [PubMed] [Google Scholar]
  32. Martikainen P, Valkonen T. Mortality after the death of a spouse: rates and causes of death in a large finnish cohort. American Journal of Public Health. 1996;86:1087–1093. doi: 10.2105/AJPH.86.8_Pt_1.1087. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Matarazzo V, Schaller F, Nédélec E, Benani A, Pénicaud L, Muscatelli F, Moyse E, Bauer S. Inactivation of Socs3 in the hypothalamus enhances the hindbrain response to endogenous satiety signals via oxytocin signaling. Journal of Neuroscience. 2012;32:17097–17107. doi: 10.1523/JNEUROSCI.1669-12.2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. McDonald PG, Antoni MH, Lutgendorf SK, Cole SW, Dhabhar FS, Sephton SE, Stefanek M, Sood AK. A biobehavioral perspective of tumor biology. Discovery Medicine. 2005;5:520–526. [PMC free article] [PubMed] [Google Scholar]
  35. Perea-Rodriguez JP, Takahashi EY, Amador TM, Hao RC, Saltzman W, Trainor BC. Effects of reproductive experience on central expression of progesterone, oestrogen α, oxytocin and vasopressin receptor mRNA in male California mice (Peromyscus californicus) Journal of Neuroendocrinology. 2015;27:245–252. doi: 10.1111/jne.12264. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Pieters M, Wolberg AS. Fibrinogen and fibrin: an illustrated review. Research and Practice in Thrombosis and Haemostasis. 2019;3:161–172. doi: 10.1002/rth2.12191. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Sahni A, Simpson-Haidaris PJ, Sahni SK, Vaday GG, Francis CW. Fibrinogen synthesized by Cancer cells augments the proliferative effect of fibroblast growth factor-2 (FGF-2) Journal of Thrombosis and Haemostasis. 2008;6:176–183. doi: 10.1111/j.1538-7836.2007.02808.x. [DOI] [PubMed] [Google Scholar]
  38. Scribner JL, Vance EA, Protter DSW, Sheeran WM, Saslow E, Cameron RT, Klein EM, Jimenez JC, Kheirbek MA, Donaldson ZR. A neuronal signature for monogamous reunion. PNAS. 2020;117:11076–11084. doi: 10.1073/pnas.1917287117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Sex GWR. Marital status mortality. AJS; American Journal of Sociology. 1973;79:45–67. doi: 10.1016/j.maturitas.2012.08.007. [DOI] [PubMed] [Google Scholar]
  40. Shimizu Y, Kinoshita I, Kikuchi J, Yamazaki K, Nishimura M, Birrer MJ, Dosaka-Akita H. Growth inhibition of non-small cell lung Cancer cells by AP-1 blockade using a cJun dominant-negative mutant. British Journal of Cancer. 2008;98:915–922. doi: 10.1038/sj.bjc.6604267. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Sloan EK, Priceman SJ, Cox BF, Yu S, Pimentel MA, Tangkanangnukul V, Arevalo JM, Morizono K, Karanikolas BD, Wu L, Sood AK, Cole SW. The sympathetic nervous system induces a metastatic switch in primary breast Cancer. Cancer Research. 2010;70:7042–7052. doi: 10.1158/0008-5472.CAN-10-0522. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Sullivan AR, Fenelon A. Patterns of widowhood mortality. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences. 2014;69:53–62. doi: 10.1093/geronb/gbt079. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Teng Y, Wang X, Wang Y, Ma D. Wnt/beta-catenin signaling regulates Cancer stem cells in lung Cancer A549 cells. Biochemical and Biophysical Research Communications. 2010;392:373–379. doi: 10.1016/j.bbrc.2010.01.028. [DOI] [PubMed] [Google Scholar]
  44. Vidman L, Källberg D, Rydén P. Cluster analysis on high dimensional RNA-seq data with applications to Cancer research - An evaluation study. PLOS One. 2019;14:e0219102. doi: 10.1371/journal.pone.0219102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Visvader JE, Lindeman GJ. Cancer stem cells: current status and evolving complexities. Cell Stem Cell. 2012;10:717–728. doi: 10.1016/j.stem.2012.05.007. [DOI] [PubMed] [Google Scholar]
  46. Westhoff B, Colaluca IN, D'Ario G, Donzelli M, Tosoni D, Volorio S, Pelosi G, Spaggiari L, Mazzarol G, Viale G, Pece S, Di Fiore PP. Alterations of the notch pathway in lung Cancer. PNAS. 2009;106:22293–22298. doi: 10.1073/pnas.0907781106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Wright EC, Parks TV, Alexander JO, Supra R, Trainor BC. Activation of kappa opioid receptors in the dorsal raphe have sex dependent effects on social behavior in California mice. Behavioural Brain Research. 2018;351:83–92. doi: 10.1016/j.bbr.2018.05.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Yuan X, Wu H, Xu H, Han N, Chu Q, Yu S, Chen Y, Wu K. Meta-analysis reveals the correlation of notch signaling with non-small cell lung Cancer progression and prognosis. Scientific Reports. 2015;5:10338. doi: 10.1038/srep10338. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Yuan L, Liu S, Bai X, Gao Y, Liu G, Wang X, Liu D, Li T, Hao A, Wang Z. Oxytocin inhibits lipopolysaccharide-induced inflammation in microglial cells and attenuates microglial activation in lipopolysaccharide-treated mice. Journal of Neuroinflammation. 2016;13:7. doi: 10.1186/s12974-016-0541-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Zhang X, Wang F, Huang Y, Ke K, Zhao B, Chen L, Liao N, Wang L, Li Q, Liu X, Wang Y, Liu J. FGG promotes migration and invasion in hepatocellular carcinoma cells through activating epithelial to mesenchymal transition. Cancer Management and Research. 2019;11:1653–1665. doi: 10.2147/CMAR.S188248. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Zhu B, Yang J-R, Jia Y, Zhang P, Shen L, Li X-L, Li J, Wang B. Overexpression of NR4A1 is associated with tumor recurrence and poor survival in non-small-cell lung carcinoma. Oncotarget. 2017;8:113977–113986. doi: 10.18632/oncotarget.23048. [DOI] [PMC free article] [PubMed] [Google Scholar] [Retracted]

Decision letter

Editor: Ivan Topisirovic1
Reviewed by: William W Lockwood2

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Acceptance summary:

This study provides evidence that pair bonding may have a favourable role in cancer. To this end, the authors show that disruption of pair bonds in monogamous rodents results in faster growth of ectopic tumors formed by various lung cancer cell lines. Moreover, evidence is provided that sera from paired and pair-disrupted animals differentially affect gene expression in cancer cells, whereby sera from pair-disrupted animals stimulate growth of cancer cells more strongly than the sera from paired animals. Overall, these findings underscore potentially significant impact of psychosomatic factors caused by changes in social environment on tumor growth rates. More broadly, these results emphasize importance of considering social environment in clinical management of neoplasia.

Decision letter after peer review:

Thank you for submitting your article "Persistent effects of pair bonding in lung tumorigenesis in monogamous rodents" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and a Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: William W Lockwood (Reviewer #2).

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Summary

In this study, Naderi et al provide evidence for beneficiary effects of pair bonding in neoplasia by showing that disruption of pair bonds in monogamous rodents accelerates growth of ectopic tumors formed by A549 lung cancer cells. Moreover, the authors provide evidence that sera from pair-disrupted animals bolster A549 spheroid growth. Overall, it was found that the observations reported in this study are of high potential interest inasmuch as they emphasize potential importance of social environment and associated psychosomatic factors (e.g. anxiety, stress) in modulating tumor growth. However, lack of mechanistic evidence linking pair bonding to tumor growth decreased the enthusiasm for the study. Some methodological limitations and issues with insufficient description of experimental procedures were also noted.

Essential Revisions:

1. It was thought that the major limitation of the study is that it relies on a single cell line (A549) which brings to question the generality of the conclusions. To this end, it was found that this study would significantly benefit from using additional cancer cell lines and/or carcinogens (to study effects of pair bonding on tumor initiation). Alternatively, the authors should clearly indicate this limitation and accordingly modify their conclusions as well as the title and abstract of the manuscript.

2. Another major issue that was observed is the apparent lack of mechanisms explaining the effects of bonding on tumor growth. Solely relying on cancer stem cell markers was found to be insufficient to render clustering analyses informative and support authors conclusion. Pertinent to this, additional factors and/or parameters of cancer growth should be included. It was suggested that performing gene expression profiling in this context may be warranted as such approach is likely to illuminate the molecular underpinnings of observed phenomena.

3. It was found that in some experiments (e.g. transplantation studies), the sample sizes were on the modest side. Sample size and statistical power information should be provided throughout the manuscript (including the abstract). Finally, it was deemed that the methodology was not sufficiently explained and that experimental details were missing. For instance, the authors should provide information of the number of replicates.

This summary was based on individual Reviewers' assessments (provided below) and subsequent discussion.

Reviewer #1:

This is an inventive paper at the intersection of the social and the biological sciences – using experimental methods in an animal model almost perfectly suited to test the ideas in question. The topic is whether the disruption of pair bonding actually can facilitate tumorigenesis, and the investigators use several in vitro and in vivo approaches, and several rodent species, to show that this is the case. The results comport with epidemiological studies that have shown that recently widowed individuals have higher risk of various diseases, including cancer.

I found this paper to be inventive, at the intersection of the social and the biological sciences, using experimental methods in an animal model, almost perfectly suited to test the ideas in question. The topic is whether the disruption of pair bonding actually can facilitate tumorigenesis, and the investigators use several approaches, comporting with epidemiological studies that have shown that recently widowed individuals have higher risk of various diseases, including cancer.

One of the things I would have liked to of seen is more attention to the temporal dynamics, including how soon after disruption of the bond there was evidence of increased tumor risk. In humans, most cancers are slow growing, and one of the issues with studies of the widowhood effect is that while it makes sense that the loss of a spouse might lead to the sudden, new onset of a stroke, say, it's hard to imagine how it can prompt the rapid emergence of colon cancer. My point is that it seems more likely that, say, a small and nascent tumor has its course accelerated, rather than initiated, by the loss of a partner. Hence, an animal model of this might involve first implanting a tumor into pair-bonded rodents, and then disrupting the pair bond for some of them and assessing how the course of illness was modified. I realize this is a different set of experiments, and I am not suggesting that these must be done at present, since the current results are themselves a very meaningful advance, in my view.

I think the abstract should include samples sizes. In addition, in the abstract, the authors might want to mention that another possible explanation for why the risk of death from cancer is higher in people who are recently widowed is the potential for homogamy based on cancer risk. In other words, imagine that people in poor health are more likely to marry people in poor health (which is true), then people with similar cancer risk would marry each other and the death of the first partner would simply be an indicator of these unknown factors predisposing to the union and then the subsequent partner would also die of cancer.

In the Results section, such as a page 4, I would include the sample size of all analyses even though the details are reported in the methods section.

Similarly, although this is mentioned in the methods section, I would make it clear at page 5 that all the mice are of the same age regardless of their mating status. We obviously want to avoid a conflation of mating status with age.

The result in the middle of page 5, which showed that, by day 25, the tumors persisted in both a bond-disrupted and bonded animals, at 89 and 82% respectively, while in virgin animals there were detectable in only 25% of cases, was quite perplexing. The authors speculate that this may have to do with concomitant anti-inflammatory action of oxytocin, but the size of the difference is so large that I was left wondering what could possibly be going on.

At page 7, I wondered how big the sample sizes were in ascertaining that certain tumors from the bonded animal groups retained a distinct profile. And, analytically, it seems the investigators clustered all three of their kinds of experiments into one hierarchical model. I'm not sure that is optimal.

Also at page 7, the paper suddenly mentioned that some of their samples were siblings. I suspect the statistical controls are adequate but I was left wondering why, methodologically, they didn't simply exclude siblings from the animals they used in their experiments?

I thought the results from the monogamous and polygamous mice were especially strong and revealing.

Overall, I thought this was a very inventive paper with many novel results that open up new frontiers in thinking about not only the underlying biosocial phenomena, but also how to experimentally manipulate and analyze them.

Reviewer #2:

Naderi et al. present an interesting short report that aims to experimentally assess the influence of pair bonding on tumorigenesis. While epidemiological studies have suggested that social interactions and the bond between couples affects tumor risk and development, this is been difficult to independently assess in humans due to potentially confounding issues associated with other lifestyle variables. Thus, the underlying biological processes regulating the observed protective effects of marriage and the adverse events correlated with the of the loss of a partner are still largely unknown. Using monogamous rodents as a model system, the authors aim to address this question and directly test the impact of pair bonding and disruption on the tumorigenic potential of human lung cancer cells. They find that sera from pair disrupted Peromyscus californicus increases the size and morphology of A549 cells grown as spheroids, while sera bonded mice had no effect compared to sera from non-pair bonded animals. Importantly, these effects were also seen with sera from P. polionotus, which is also monogamous, but not from the polygamous P. manicultus where no effects of pair disruption were observed. This effect was also demonstrated in vivo when A549 cells were transplanted into pair bonded, disrupted or virgin animals or from these animals into nude mice, suggesting the long-lasting effects on tumorigencity of the cells.

Overall, the study is innovative for its use of monogamous rodent species for the study of the long outstanding question of the biological effects of social interactions on cancer development. As the most common model organism for the study of cancer, Mus musculus, is promiscuous and does not form long-term pair bonds, these results suggest that important factors regulating cancer development may be lacking in many preclinical studies. This is an important consideration and may influence how studies are modeled in the future to capture additional host aspects associated with tumorigenesis. The authors ultimately succeed in demonstrating that pair bonding should be considered in this regard as their in vitro experiments, while limited in breadth, do indicate clear effect of sera from pair disrupted animals on tumorigenic potential across monogamous, and not polygamous, species. While the authors do not comprehensively assess the underlying processes regulating these phenotypes, they demonstrate that cells/tumors from pair disrupted animals cluster together based on the expression of a limited set of cancer stem cell markers. Together, the authors present sufficient in vitro and in vivo data to conclude that pair-bonding affects tumor cell behaviour.

While unique in design and informative in its overall conclusions, it should be noted that the study does have many important limitations. First, only a single lung cancer cell line (A549) is used in the study and it is difficult to determine whether the observed findings will be generalizable across different tumor types (aside from lung adenocarcinoma) or those with different mutation spectra and driver oncogenes (aside from KRAS G12S). Furthermore, limited mice are used in transplantation experiments and it is also unclear whether there is substantial diversity in the effects from sera from different animals. The number of animals used to isolate sera is not indicated and therefore, it is unclear how variable the effects on spheroid growth are and whether this influences the conclusions.

Another limitation is the use of cancer cell lines. These cells are already tumorigenic and pair disruption in this instance is therefore testing whether there is an influence on making these cells more or less tumorigenic. This is different than the real-life scenario where it is likely that spouses of deceased partners will not have cancer in the first place, but instead, may develop cancer as a result of the stress and other factors associated with pair disruption.

The assessment of only select cancer stem cell markers is an additional limitation. While these are no-doubt important for influencing tumorigenic potential, they represent only a fraction of the changes that could be occurring as a result of pair bonding disruption. As such, this analysis is very limited in breadth. Furthermore, it is difficult to assess the clustering results as the limited variables tested impacts the ability to discriminate between samples. Thus, this type of analysis isn't the most appropriate as currently presented and will likely require the assessment of additional variables to accurately separate the samples by underlying phenotypes.

Another limitation is the use of cancer cell lines. These cells are already tumorigenic and the influence of pair disruption in this instance is therefore testing whether it can make these cells more or less tumorigenic. This is different than the real-life scenario where it is likely that spouses of deceased partners will not have cancer in the first place, but instead, may develop cancer as a result of the stress and other factors associated with pair disruption. Thus, a more appropriate and complementary model would assess the effects of pair disruption on cancer initiation as opposed to development and progression. For example, pair bonded and disrupted animals could be given a carcinogen to assess the rate and incidence of tumor onset. Or, cell lines that represent a pre-malignant state (ie. don't form tumors in nude mice) could be used to assess the effects on progressing these cells to a fully malignant state.

Reviewer #3:

The work provides an interesting observation that the bonding between monogamous rodents may help in diminishing the ectopically inoculated tumour size and a disruption of this pair bonds resulted in increased tumour size probably due to anxiety and stress. This paper will be of interest to support psychosomatic medicine in cancer treatments. The authors used sera collected from three groups (virgin, bonding, and disruption of bonding) monogamous rodents to evaluate the effect on tumour cells growing in spheroids. Tumour cells were also subcutaneously inoculated into flank of the three groups of mice to evaluate the actual tumour growth.

Strengths:

An interesting approach to use both tumour cells grown on spheroids and in mice with different pairing bonding conditions. It was relatively convincing results that tumour cells grow faster in a condition associated with bonding disruption.

Weaknesses:

Although the paper has strengths in principle, the weaknesses of the paper are that these strengths are only reflected in a descriptive manner. Epidemiological evidence already suggests that social interactions and especially bonding between couples influence tumorigenesis. The observation in this paper is interesting and confirmatory of cancer patients but not novel. It would have been informative had the authors utilised the models to perform a more in-depth investigation in providing mechanistic insight on how the bonding disruption affects tumour growth. Some experimental data and experimental details require clarification to allow confident interpretation of the described results. The brief discussion was presented mostly as a summary and statement without sufficient literature support and citation, which should be expanded and enhanced.

The manuscript submitted by Naderi et al. entitled "Persistent effects of pair bonding in lung tumorigenesis in monogamous rodents" includes interesting observations that pair-bonding between monogamous rodents (P. californicus) is beneficial to the mice in preventing the implanted tumor growth. However, the current version of the submitted manuscript includes the only limited scope and lack of mechanistic insight. Multiple experimental details were missing that prevent a confident interpretation of the described results.

1. In figure 1b, it is not clear when the serum from the bonding (B) group was collected (only one B condition was shown). The appropriate control should include serum from the bonding group at all time points (24h, 1week, and 2 weeks) as the bonding disruption (BD) group for a direct comparison.

2. There is a mistake in figure 2b. How could it be possible that the 6 mice with the measurable tumor in the virgin (V) group only accounted for 25% of the tumor-bearing sub-group in the pie chart when the total mice number was 8? Based on the results in Figure 2a, the mouse number with a tumor should be 6 and accounted for 75% (6/8) of the total virgin group. Page 5, sentence 99-105 should be better included in the discussion than in the Results section. In addition, the difference is only between 82% (B) and 75% (V), rather than the described 25%, which did not seem to be significant.

3. In figure 2d, how many tumors were transplanted? How many nude mice were used? The information was not included in the methods, nor in the results or figure legends. All of the "N" should be clearly described in the figure legends. Were all tumor cells isolated from all 9 tumor-bearing mice in the B group and 8 tumor-bearing mice from the bonding disruption (BD) group dissected and used for transplantation study? How many different tumor cells from each group (V, B, and BD) were transplanted into nude mice? How many nude mice were transplanted with tumor cells from each tumor-bearing mouse?

4. It is not clear how many times the in vivo animal experiments were performed? Did the in vivo data showing in figure 2 represent only a one-time experiment?

5. In figure 3, the signature using only three genes was overly simplified. The selected markers did not provide supporting information on why serum from the different mice groups displayed various sizes in spheroids and tumors because the gene expression of the BD groups was consistently located in between B and V groups, which all showed smaller tumor sizes.

6. The clustering description of figure 3 was confusing and the signature was not unique, which was not unexpected because the expression data of these genes did not significantly differ among the different groups. Non-biased analysis using sequencing results with unsupervised clustering likely will provide data that are more informative and mechanistic.

7. The sentence from lines 127-130 is an over-interpretation of the results. The description of transcription could be better included in the discussion and not in the Results section.

8. In figure 4, the description of the serum collection was absent. When was the sera collection time? Were the polygamous mice housed individually after disruption (probably yes) and separated from all-female mice. The figure legend was incomplete and lack important details. The abbreviation should also be described.

9. The data in figure 4 regarding the effects on different shapes of spheroids (but without clear biological consequence) did not positively support the hypothesis. Additional mouse species may be needed to identify if the tumor size difference observed in P. californicus after bonding disruption was an exception or a general observation. Alternatively, the authors should limit the claim and re-write the title to reflect the observation in the specific mouse strain.

10. In figure 4, the serum collected from the "same animal" at bonding and following the bond disruption did not take into consideration of aging and hormonal effects. Additional groups of animals should have been included to address these concerns and avoid the over-interpretation of the results.

11. The brief discussion was presented mostly as a summary and statement without sufficient literature support and citation, which should be expanded and enhanced.

eLife. 2021 May 7;10:e64711. doi: 10.7554/eLife.64711.sa2

Author response

Essential Revisions:

1. It was thought that the major limitation of the study is that it relies on a single cell line (A549) which brings to question the generality of the conclusions. To this end, it was found that this study would significantly benefit from using additional cancer cell lines and/or carcinogens (to study effects of pair bonding on tumor initiation). Alternatively, the authors should clearly indicate this limitation and accordingly modify their conclusions as well as the title and abstract of the manuscript.

We agree with this comment of the reviewers and editors. To address this, we used additional lung cancer cell lines (H1703, H596, H358, and H292). In the absence of animals used originally that were 14-17 months old and bonded for 1 year or more, younger animals (6-8 months old) bonded for 2 months were now used. Not only we were able to record for all cells lines similar effects with analogous variation as to that recorded originally (now shown as scatter dot plots in original Figure 1c), but more importantly we think, we were able to show that the effect varies among individuals and is produced in the different lung cancer cells in a similar manner (Figure 1—figure supplement 1).

2. Another major issue that was observed is the apparent lack of mechanisms explaining the effects of bonding on tumor growth. Solely relying on cancer stem cell markers was found to be insufficient to render clustering analyses informative and support authors conclusion. Pertinent to this, additional factors and/or parameters of cancer growth should be included. It was suggested that performing gene expression profiling in this context may be warranted as such approach is likely to illuminate the molecular underpinnings of observed phenomena.

We agree with the reviewers that the qPCR data on selected cancer stem cell markers are insufficient to support the clustering findings we described in the original manuscript. In this revised version of our manuscript, we included whole transcriptome RNAseq results that confirmed that cells cluster together, depending on the bonding history of the serum donors. In addition, differential expression analysis was performed which pointed to bonding-dependent effects in cell adhesion and migration as well as tissue morphogenetic processes. In the revised version of the manuscript, to accommodate these new results, extensive changes were performed in the abstract, the discussion, the Results section and the figures. We also moved the original qPCR data of cancer stem cell markers into the supplementary information which now, also includes the new clustering data from RNAseq (Figure3 and Figure 3—figure supplement 2,3).

3. It was found that in some experiments (e.g. transplantation studies), the sample sizes were on the modest side. Sample size and statistical power information should be provided throughout the manuscript (including the abstract). Finally, it was deemed that the methodology was not sufficiently explained and that experimental details were missing. For instance, the authors should provide information of the number of replicates

In compliance with the reviewers’ comments details about statistical power, number of replicas and animal numbers in tumor transplantation studies are now mentioned throughout the revised manuscript.

This summary was based on individual Reviewers' assessments (provided below) and subsequent discussion.

Reviewer #1:

[…] One of the things I would have liked to of seen is more attention to the temporal dynamics, including how soon after disruption of the bond there was evidence of increased tumor risk. In humans, most cancers are slow growing, and one of the issues with studies of the widowhood effect is that while it makes sense that the loss of a spouse might lead to the sudden, new onset of a stroke, say, it's hard to imagine how it can prompt the rapid emergence of colon cancer. My point is that it seems more likely that, say, a small and nascent tumor has its course accelerated, rather than initiated, by the loss of a partner. Hence, an animal model of this might involve first implanting a tumor into pair-bonded rodents, and then disrupting the pair bond for some of them and assessing how the course of illness was modified. I realize this is a different set of experiments, and I am not suggesting that these must be done at present, since the current results are themselves a very meaningful advance, in my view.

We completely agree with the reviewer that indeed, bond disruption most likely accelerates the development of a preexisting condition, rather than initiating it. This is also supported by our model at which the cancer cells were implanted in the animals, just prior to the disruption of the pair bonds. This point is now discussed in the Discussion by the addition of the following sentence:

Discussion (2nd paragraph): “…In our animal model, cancer cells were implanted in tumor-free animals and the kinetics of tumorigenesis were affected by the animals’ bonding history. Whether pair bonding and disruption can also influence tumor initiation will have to be established, nevertheless, the fact that most cancers are slow growing in patients is consistent with effects of widowhood in influencing the progression, as opposed to the initiation, of the disease.”

The temporal dynamics point, is also of major significance. Some hints related to this are provided in the spheroid assays at which sera collected at different time points are used. We have initiated such studies to describe in higher detail these dynamics, but the genetic heterogeneity of the deer mice requires high numbers of animals to accurately describe these differences. We hope that we will be able to describe them in a future publication. In the meanwhile, however we included in this paper as new Figure 1—figure supplement 1, the results of a study involving 7 younger (about 6-8 months old as opposed to the original study that involved 14-17 months old animals) that were bonded for 2 months (as opposed to 12 months of the original study) or bond-disrupted. In this study we found similar effects in spheroid formation with those reported originally. More importantly we obtained evidence that the effect varies among individuals but persists across different lung cancer cells.

I think the abstract should include samples sizes. In addition, in the abstract, the authors might want to mention that another possible explanation for why the risk of death from cancer is higher in people who are recently widowed is the potential for homogamy based on cancer risk. In other words, imagine that people in poor health are more likely to marry people in poor health (which is true), then people with similar cancer risk would marry each other and the death of the first partner would simply be an indicator of these unknown factors predisposing to the union and then the subsequent partner would also die of cancer.

We thank the reviewer for his/her insightful comment. Sample sizes for tumor re-transplantation are now shown in the abstract. The possibility of homogamy as an additional factor explaining the widowhood effect is now described in the abstract.

Abstract (line 2): “…lifestyle changes, homogamy (likelihood of individuals to marry people of similar health) or directly associated with…”

In the Results section, such as a page 4, I would include the sample size of all analyses even though the details are reported in the methods section.

Similarly, although this is mentioned in the methods section, I would make it clear at page 5 that all the mice are of the same age regardless of their mating status. We obviously want to avoid a conflation of mating status with age.

Sample sizes and the age of the animals are now included in the Results section of the revised manuscript.

The result in the middle of page 5, which showed that, by day 25, the tumors persisted in both a bond-disrupted and bonded animals, at 89 and 82% respectively, while in virgin animals there were detectable in only 25% of cases, was quite perplexing. The authors speculate that this may have to do with concomitant anti-inflammatory action of oxytocin, but the size of the difference is so large that I was left wondering what could possibly be going on.

We agree with the reviewer that this is really puzzling. We moved the corresponding part in the discussion (paragraph 3) and also added the following at the end:

“…It is noted though that the high difference in the tumorigenicity between virgins and the bonded or bond-disrupted animals, renders differential immune suppression unlikely as the sole contributor for this discrepancy.”

At page 7, I wondered how big the sample sizes were in ascertaining that certain tumors from the bonded animal groups retained a distinct profile. And, analytically, it seems the investigators clustered all three of their kinds of experiments into one hierarchical model. I'm not sure that is optimal.

We agree that clustering analysis, especially as described in the original version of our manuscript suffers from various limitations. Therefore, it is not discussed in detail and the results are moved in the supplementary material (Figure 3-Figure Suppl. 1) in the revised manuscript. However, unsupervised clustering was also performed with the RNAseq data which showed similar trends (Figure 3-Figure Suppl. 2). This analysis showed that expression profiles of bonding and disruption are more robust than those of virgins, as discussed in the revised manuscript (Author response image 1).

Author response image 1. Expression signature of human A549 lung cancer cells cultured in sera from deer mice differing in bonding history.

Author response image 1.

Open in a new tab

We refrained however from describing these results explicitly here, as various analyses are still ongoing and their description is deviating from the original scope of the present study. We wanted to note that by using these RNAseq data (n=6 individuals per group) we were able to define a bonding signature consisting of 15 genes that showed that the discriminatory effects of bonding are more potent than those of disruption (bond-disrupted animals cluster together with virgins while bonded deviate earlier). More surprisingly, this signature predicted lung cancer survival of human patients (Author response image 2).

Author response image 2. Bonding signature and patients’ prognosis.

Author response image 2.

Open in a new tab

a. Hierarchical clustering of lung cancer patients from TCGA based on the 13 genes signature and survival. b. Survival probability of the two groups of TCGA lung cancer patients that emerged by using the 13 gene signature.

Also at page 7, the paper suddenly mentioned that some of their samples were siblings. I suspect the statistical controls are adequate but I was left wondering why, methodologically, they didn't simply exclude siblings from the animals they used in their experiments?

The animals are genetically diverse (outbred) and a strategy we occasionally use to minimize potential founder effects is to use siblings distributed in different groups. We tried to do this here but unfortunately, tumor take rates and availability of animas did not allow us to do this rigorously. To avoid confusion and unjustified speculations we now eliminated this part of the discussion (along with the related discussion of clustering) and mentioned in the Methods the sporadic use of siblings (pointing to Figure 1-Figure Suppl. 1) at which mating numbers of the crosses are shown.

I thought the results from the monogamous and polygamous mice were especially strong and revealing.

Overall, I thought this was a very inventive paper with many novel results that open up new frontiers in thinking about not only the underlying biosocial phenomena, but also how to experimentally manipulate and analyze them.

Reviewer #2:

[…] Weaknesses:

While unique in design and informative in its overall conclusions, it should be noted that the study does have many important limitations. First, only a single lung cancer cell line (A549) is used in the study and it is difficult to determine whether the observed findings will be generalizable across different tumor types (aside from lung adenocarcinoma) or those with different mutation spectra and driver oncogenes (aside from KRAS G12S). Furthermore, limited mice are used in transplantation experiments and it is also unclear whether there is substantial diversity in the effects from sera from different animals. The number of animals used to isolate sera is not indicated and therefore, it is unclear how variable the effects on spheroid growth are and whether this influences the conclusions.

We agree with the reviewer for the various limitations he pointed out. In this revised version of the manuscript these points were addressed. Specifically, additional lung cancer cell lines were used to explore how general our findings are. The diversity point of the experimental animals, indeed represents an advantage but also a disadvantage of this model. Although randomly animals were assigned into groups, founder effects remain formally possible. Sample numbers, besides legends are also now mentioned throughout the manuscript as well.

Another limitation is the use of cancer cell lines. These cells are already tumorigenic and pair disruption in this instance is therefore testing whether there is an influence on making these cells more or less tumorigenic. This is different than the real-life scenario where it is likely that spouses of deceased partners will not have cancer in the first place, but instead, may develop cancer as a result of the stress and other factors associated with pair disruption.

This is a very important point that we tried to address in the revised manuscript by adding a relevant comment in the Discussion (2nd paragraph). See also response to reviewer 1. Briefly, we agree that our experimental set up addresses the effects of bonding and disruption in preexisting cancers that now become more aggressive after the disruption of pair bonds.

The assessment of only select cancer stem cell markers is an additional limitation. While these are no-doubt important for influencing tumorigenic potential, they represent only a fraction of the changes that could be occurring as a result of pair bonding disruption. As such, this analysis is very limited in breadth. Furthermore, it is difficult to assess the clustering results as the limited variables tested impacts the ability to discriminate between samples. Thus, this type of analysis isn't the most appropriate as currently presented and will likely require the assessment of additional variables to accurately separate the samples by underlying phenotypes.

As we already mentioned we agree that clustering analysis with these 3 stem cell markers may be misleading. Therefore, the description of these data was moved to the supplementary information and extensive discussion on this was avoided. Nevertheless, RNAseq analysis was performed and clustering according to whole transcriptome data, indicated similar trends. These new data were also mentioned in the revised manuscript and are shown as a new Figure 3-Figure Suppl. 2.

Another limitation is the use of cancer cell lines. These cells are already tumorigenic and the influence of pair disruption in this instance is therefore testing whether it can make these cells more or less tumorigenic. This is different than the real-life scenario where it is likely that spouses of deceased partners will not have cancer in the first place, but instead, may develop cancer as a result of the stress and other factors associated with pair disruption. Thus, a more appropriate and complementary model would assess the effects of pair disruption on cancer initiation as opposed to development and progression. For example, pair bonded and disrupted animals could be given a carcinogen to assess the rate and incidence of tumor onset. Or, cell lines that represent a pre-malignant state (ie. don't form tumors in nude mice) could be used to assess the effects on progressing these cells to a fully malignant state.

As discussed above, our findings only explain one aspect of the phenomenon, the effect of widowhood in cancer progression. Whether this also influences cancer initiation, while highly relevant and of particular interest, as the reviewer recognizes cannot be assessed by this model. We are exploring feasible options to address this in our lab, including those suggested by the reviewer and we thank him for that (chemical carcinogenesis and premalignant lesions transplantation).

Reviewer #3:

[…] Weaknesses:

Although the paper has strengths in principle, the weaknesses of the paper are that these strengths are only reflected in a descriptive manner. Epidemiological evidence already suggests that social interactions and especially bonding between couples influence tumorigenesis. The observation in this paper is interesting and confirmatory of cancer patients but not novel. It would have been informative had the authors utilised the models to perform a more in-depth investigation in providing mechanistic insight on how the bonding disruption affects tumour growth. Some experimental data and experimental details require clarification to allow confident interpretation of the described results. The brief discussion was presented mostly as a summary and statement without sufficient literature support and citation, which should be expanded and enhanced.

The main point of this first study was to address if the well-described, by the epidemiological studies, widowhood effect has biological basis, besides the confounding effects of lifestyle changes and homogamy. We think that our results will now pave the way for the implementation of mechanistic studies addressing this phenomenon. Some mechanistic insights though we believe are provided by the inclusion of RNAseq studies in the revised manuscript.

The manuscript submitted by Naderi et al. entitled "Persistent effects of pair bonding in lung tumorigenesis in monogamous rodents" includes interesting observations that pair-bonding between monogamous rodents (P. californicus) is beneficial to the mice in preventing the implanted tumor growth. However, the current version of the submitted manuscript includes the only limited scope and lack of mechanistic insight. Multiple experimental details were missing that prevent a confident interpretation of the described results.

1. In figure 1b, it is not clear when the serum from the bonding (B) group was collected (only one B condition was shown). The appropriate control should include serum from the bonding group at all time points (24h, 1week, and 2 weeks) as the bonding disruption (BD) group for a direct comparison.

We agree with the reviewer’s point about a detailed evaluation of the temporal dynamics of bonding and disruption. In this study we wanted to simulate widowhood and therefore we used a setting of extended bonding, followed by a “small-scale” time course of disruption (1day, 1 week and 2 weeks). Nevertheless, such detailed assessment is in our future plans. In addition, more details about the experimental conditions are now included in the Results, besides the Methods section of the manuscript.

2. There is a mistake in figure 2b. How could it be possible that the 6 mice with the measurable tumor in the virgin (V) group only accounted for 25% of the tumor-bearing sub-group in the pie chart when the total mice number was 8? Based on the results in Figure 2a, the mouse number with a tumor should be 6 and accounted for 75% (6/8) of the total virgin group. Page 5, sentence 99-105 should be better included in the discussion than in the Results section. In addition, the difference is only between 82% (B) and 75% (V), rather than the described 25%, which did not seem to be significant.

We apologize for the confusion in the description of the tumorigenicity results which was due to the fact that the numbers referred to different time points (day 15 and 25, at which several virgins lost their tumors). This point is now better clarified in more detail in the revised manuscript (and the corresponding figure). Per the reviewer’s suggestion the oxytocin-related discussion was transferred from the results to the Discussion section.

3. In figure 2d, how many tumors were transplanted? How many nude mice were used? The information was not included in the methods, nor in the results or figure legends. All of the "N" should be clearly described in the figure legends. Were all tumor cells isolated from all 9 tumor-bearing mice in the B group and 8 tumor-bearing mice from the bonding disruption (BD) group dissected and used for transplantation study? How many different tumor cells from each group (V, B, and BD) were transplanted into nude mice? How many nude mice were transplanted with tumor cells from each tumor-bearing mouse?

We thank the reviewer for pointing these issues that required clarification. All these details are now described in the Results section. In addition (n), besides the results are also described in the figure legend. As now described in the manuscript each Peromyscus-grown tumor was implanted in a single nude mouse.

4. It is not clear how many times the in vivo animal experiments were performed? Did the in vivo data showing in figure 2 represent only a one-time experiment?

Actually, due to the limited availability of deer mice at a given time point, the experiment was performed serially, in 2-3 phases and results were pooled together: As soon as mice became available they were distributed in groups and the experiment was performed. Unfortunately, these are outbred animals at which birth of pups in the colony does not occur simultaneously to allow full synchronization of all experiments. However, as such, we feel that the chances for reflecting a “one-time” observation due to methodological issues, are limited.

5. In figure 3, the signature using only three genes was overly simplified. The selected markers did not provide supporting information on why serum from the different mice groups displayed various sizes in spheroids and tumors because the gene expression of the BD groups was consistently located in between B and V groups, which all showed smaller tumor sizes.

6. The clustering description of figure 3 was confusing and the signature was not unique, which was not unexpected because the expression data of these genes did not significantly differ among the different groups. Non-biased analysis using sequencing results with unsupervised clustering likely will provide data that are more informative and mechanistic.

We fully agree with these 2 comments (#5 and #6) and therefore, in the revised manuscript, as also described in our responses to the 1st and 2nd reviewer, the discussion of clustering data, based on CSC markers, were “scaled down” and added in the supplementary information. In addition, as the reviewer suggested, we performed RNA sequencing and the results corroborated our original observations on the clustering based on bonding experiences. Importantly, these analyses provided insights regarding the potential mechanisms, pointing to effects in cell migration and adhesion, and tissue morphogenesis. Furthermore, they indicated that clustering occurs by using whole transcriptome RNAseq data and is more intense between animals (sera) from the bonded and bond-disrupted groups, while virgins have a less “rigid” behavior.

7. The sentence from lines 127-130 is an over-interpretation of the results. The description of transcription could be better included in the discussion and not in the Results section.

We agree. The corresponding description was changed to avoid potential over-interpretation of the results and was included in the Discussion, based primarily on the whole transcriptome data (Discussion, paragraph 5).

8. In figure 4, the description of the serum collection was absent. When was the sera collection time? Were the polygamous mice housed individually after disruption (probably yes) and separated from all-female mice. The figure legend was incomplete and lack important details. The abbreviation should also be described.

We thank the reviewer for pointing this out. We have now included all this missing information in the legend and the Results section.

9. The data in figure 4 regarding the effects on different shapes of spheroids (but without clear biological consequence) did not positively support the hypothesis. Additional mouse species may be needed to identify if the tumor size difference observed in P. californicus after bonding disruption was an exception or a general observation. Alternatively, the authors should limit the claim and re-write the title to reflect the observation in the specific mouse strain.

We agree that the fact that disruption affected differently the spheroids in the two monogamous species was somewhat puzzling. Thus, we did not insist on interpreting the actual effect but rather its occurrence, in the monogamous species only. We note though that it is always possible that this is actually the same effect but recorded at different stages (cell dispersion seen in initially in maniculatus and later, after cells proliferated in californicus as well). It is conceivable that a rigorous time course and dose-response experiment may be able to address all these differences. While these are in our future plans, at this point we feel that may be sufficient to report our observations by emphasizing that an effect was seen. To clarify this we also added in the corresponding description the following:

(Results, 6 lines before the end): “…In addition, it may reflect the same effect (cell dispersion followed by proliferation) but recorded at different stages during the formation of the spheroids.”

In addition, we repeated the experiment and performed statistical analysis on the frequency of its occurrence (n=12 animals for PO and 12 for BW; P=0.0004, chi-sq test; Figure 4-Figure Suppl. 1) which also indicated that P. maniculatus occasionally exhibit the “scattered” phenotype at bonding as well.

Finally, the title was changed, as suggested.

10. In figure 4, the serum collected from the "same animal" at bonding and following the bond disruption did not take into consideration of aging and hormonal effects. Additional groups of animals should have been included to address these concerns and avoid the over-interpretation of the results.

We thank the reviewer for this comment and we have to mention that actually, we believe that hormonal effects of bonding and bond disruption indeed constitute the mechanistic basis for our findings. As regards the age effects, although it remains formally plausible, for animals that are about 1.5 years old (at bonding when serum samples were obtained), serum sampling 1- 2 weeks later appears unlikely to reflect age-related effects. This was also addressed in the B-BD comparison of the results in new Figure 1-Figure Suppl. 1 at which different animals (but sibling pairs) were used.

11. The brief discussion was presented mostly as a summary and statement without sufficient literature support and citation, which should be expanded and enhanced.

We agree with the reviewer that the discussion of our original version of the manuscript was quite descriptive and not very critical and insightful. To some extent this was due to the fact that we wanted to refrain from overestimating some findings and engage in speculations. Nevertheless, in this revised version of our paper the Discussion section was significantly changed and a more critical approach was applied.

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Data Citations

    1. Naderi A, Ji H, Shtutman M, Kiaris H. 2021. Whole Transcriptome RNA-seq from A549 cells exposed to P. californicus serum. NCBI Gene Expression Omnibus. GSE167827

    Supplementary Materials

    Transparent reporting form
    elife-64711-transrepform.docx (245.9KB, docx)

    Data Availability Statement

    Peromyscus animals are available from the Peromyscus Genetic Stock Center.

    The following dataset was generated:

    Naderi A, Ji H, Shtutman M, Kiaris H. 2021. Whole Transcriptome RNA-seq from A549 cells exposed to P. californicus serum. NCBI Gene Expression Omnibus. GSE167827

    Articles from eLife are provided here courtesy of eLife Sciences Publications, Ltd

    RESOURCES