We experimentally controlled magnetic field exposure during planarian regeneration using a custom μ-metal enclosure (MagShield Box) to block external fields combined with Helmholtz coils to produce uniform magnetic fields at specific strengths (Figure 2). To test the hypothesis that different WMFs will produce opposite effects on new tissue growth that occur largely through modulation of radical formation, we examined both ROS accumulation and blastema formation following exposure to a controlled range of WMFs from 0 μT to 900 μT, in 100 μT increments (Figure 3). Controls were exposed to an Earth-normal 45 μT WMF, similar in strength to the geomagnetic field (which ranges from 25–65 μT). Planarian trunk fragments were created by transverse amputation just above and below the pharynx (removing both the head and tail) and regeneration was assessed at the anterior wound site. Trunk fragments were exposed to specific WMF strengths within 5 min of amputation and thereafter until analysis. The radical pair mechanism predicts that we should see some field strengths that increase as well as those that decrease ROS levels and regenerative growth.

ROS accumulation was assessed at 1 h after injury, when it has been shown that ROS is upregulated at the wound site [15]. To visualize ROS levels in live regenerates, we used a general oxidative stress indicator dye (CM-H₂DCFDA) that fluoresces upon ROS activity (Figure 3A). This allowed for the quantification of signal intensities and the statistical comparison of ROS accumulation at each field strength (Figure 3B). Our results show that compared to 45 μT controls, exposure to 200 μT WMFs prevented injury-induced ROS accumulation, while exposure to 400, 500, and 900 μT WMF exposures all caused significant increases in ROS levels. The greatest WMF effects were seen at 200 μT for inhibition and 500 μT for increased ROS accumulation.

Subsequent new tissue growth was assessed at 3 days after amputation, when blastema formation is considered complete [41]. The blastema is easily recognizable at this stage as white tissue at the wound site, since pigmentation has not yet occurred (Figure 3C). To account for any differences in worm size, blastema size was calculated as a percentage of total regenerate size (Figure 3D). We found that 100–400 μT exposures decreased blastema size, whereas at both 500 and 900 μT we observed the formation of larger than normal blastemas. Similar to our ROS findings, the greatest WMF effects on new tissue growth were seen at 200 μT for inhibition and 500 μT for increased blastema size. Our results demonstrate that WMFs can either increase or decrease both wound site ROS levels and tissue growth in a field strength-dependent manner.

These data suggest that a threshold potential exists to modulate tissue growth through WMF manipulation of ROS formation. Furthermore, they support 1) the hypothesis that WMF effects are consistent with our theoretical model, and 2) our hypothesis that WMF effects result mainly from the manipulation of ROS signaling. If correct, then we can predict equal and opposite changes in events mediated by ROS signaling, which in planarians includes control of adult stem cell behaviors after injury. Therefore, we next examined the effects of 200 and 500 μT WMFs (as representative of our observed effects) on ROS signaling and the resulting behaviors of stem cells during regeneration (Figure 4).

At 3 days after amputation, we investigated the expression of the chaperone heat shock protein 70 (Hsp70) (Figure 4A, top panels), which is involved in stress responses and cell survival [42, 43]. In planarians, blastema-associated Hsp70 expression requires injury-induced ROS, and in turn Hsp70 upregulation is required for ROS-mediated stem cell responses during regeneration [15, 44]. Therefore, at the same time point we also looked at the stem cell population using the general stem cell marker Piwi-1, as well as the late stem cell progeny marker Agat-1 (Figure 4A, middle panels). Our data showed that as predicted, as compared to controls, 200 μT WMFs caused a significant reduction in the expression of all three genes at the wound site, while 500 μT WMFs significantly increased expression (Figure 4B). These data demonstrate that WMF exposure can be used to directly inhibit or activate ROS signaling, depending on field strength.

Furthermore, we investigated WMF effects on proliferation at 3 h after injury (Figure 4A, bottom panels). In planarians, stem cells have been found to be the only actively dividing cell population. Thus, we examined stem cell proliferation by looking at the presence of phosphorylated Histone H3 (pH3), which labels mitotically active cells. We found that at 200 μT there were significantly fewer mitotic cells, while at 500 μT there was a significant increase in the number of mitotic cells (Figure 4B). These results are consistent with our prediction that WMFs could both inhibit the activation of stem cell proliferation following injury as well as increase the proliferative response.

Together, our data indicate that exposure to WMFs produces non-stochastic changes that are predictable based on our theoretical principles (Figure 1A), which suggest that different field strengths have opposing effects. Furthermore, the data provide strong evidence that WMF effects on proliferation and tissue growth are consistent with the manipulation of ROS. These results support further investigation into the potential use of WMFs as a tool to alter stem cell activity.

The majority of cellular ROS signaling is transduced by either H₂O₂ or O₂ ⁻ [45, 46]. In planarians, both H₂O₂ and O₂ ⁻ are present at the wound site following injury [47]. Therefore, we next sought to examine the effects of WMF exposures on these specific species (Figure 5). We hypothesized that WMFs modulate ROS signaling by influencing the formation of H₂O₂, since it has been well demonstrated as an ROS mediator of traditional signaling pathways.

To test this, we exposed regenerating planarians to 200 and 500 μT WMFs (with 45 μT controls) as before, and then examined the levels of H₂O₂ using the species-specific fluorescent reporter dye peroxy orange 1. Since with our general ROS indicator dye (Figure 3) we observed a peak at 1 h after injury, we chose that time point to examine H₂O₂ levels at the wound site (Figure 5A). However, there were no significant changes in the amount of H₂O₂ at either 200 or 500 μT (Figure 5B). In case there was a time delay in WMF effects specifically on H₂O₂, we also tested for effects at 2 h after injury but did not observe any significant changes (Figures 5A, B).

We then repeated these same analyses with the O₂ ⁻ specific fluorescent reporter dye orange 1. The data show that WMFs do alter wound site O₂ ⁻ levels at both 1 and 2 h after injury (Figure 5C). At 1 h, 200 μT WMF exposure significantly reduced O₂ ⁻ accumulation, although 500 μT produced no change; while at 2 h, 200 μT reduced and 500 μT increased levels of O₂ ⁻ at the wound site (Figure 5D). This pattern of opposing inhibition and activation of O₂ ⁻ by WMFs mirrored our results seen from WMF effects on ROS-mediated stem cell activity (Figure 4). However, these results were inconsistent with our hypothesized role for H₂O₂ in mediating WMF effects during tissue growth.

Studies suggest that H₂O₂ signaling plays a role during planarian regeneration [15, 39, 40, 47]. Therefore, we further investigated the possible differential roles for H₂O₂ and O₂ ⁻ in mediating the effects of WMFs on planarian regeneration (Figure 6). The general flavoenzyme inhibitor diphenyleneiodonium chloride (DPI) is often used as a pharmacological NADPH oxidase inhibitor [48, 49]. To confirm a role for H₂O₂ during regeneration, we examined the ability of exogenous H₂O₂ (which is cell permeable and readily diffuses across the plasma membrane) to rescue tissue growth following general ROS inhibition by DPI (Figures 6A–C). We pre-exposed animals to either 10 μM DPI or its vehicle control dimethyl sulfoxide (DMSO), amputated to produce trunk fragments, then allowed fragments to regenerate without drug exposure. At 3 days after amputation, blastema formation was significantly inhibited, while the addition of 400 μM H₂O₂ after amputation was able to rescue/overcome this chemical block of ROS (Figure 6C).

We repeated this H₂O₂ rescue assay but following inhibition of ROS by 200 μT WMF exposure, and without any pre-exposure before amputation (Figures 6D–F). Unlike chemical ROS inhibition, we found that WMF inhibition of blastema formation at 3 days could not be rescued by the addition of H₂O₂ (Figure 6F). To further support these findings, we also analyzed the effects of exogenous H₂O₂ on O₂ ⁻ levels without experimentally controlled WMF exposure (Figures 6G–I). Adding H₂O₂ alone, even with adding a 24 h pre-treatment, failed to significantly affect injury-induced O₂ ⁻ levels at the wound site at 2 h after injury (Figure 6I). As our results reveal that exposure to WMFs was able to alter the injury-induced accumulation of O₂ ⁻ at this same time point, the data suggest that WMF effects on tissue growth do not occur via H₂O₂, but instead are O₂ ⁻—mediated.

During our investigation into the effects of specific WMFs on H₂O₂ and O₂ ⁻ levels at the wound site, we noticed there appeared to be a difference in levels of individual species accumulation at 1 h versus 2 h. Furthermore, in these WMF experiments the apparent pattern of peak species accumulation seemed to differ between H₂O₂ and O₂ ⁻ (Figure 5). To better probe the temporal kinetics of ROS accumulation without confounding variables, we investigated normal H₂O₂ and O₂ ⁻ levels during tissue growth without experimentally controlled WMF exposure (Figure 7). Our data show that peak H₂O₂ accumulation after injury occurred at 1 h, with a significant decrease by 2 h. Conversely, while O₂ ⁻ was present at the wound site by 1 h, O₂ ⁻ levels did not peak until 2 h.

During metabolism, O₂ ⁻ is converted into H₂O₂ by the enzyme superoxide dismutase (SOD), which effectively increases H₂O₂ levels by reducing O₂ ⁻ levels (Figure 1B). If this enzymatic pathway is a main driver of ROS signaling during tissue growth, then we would predict that O₂ ⁻ accumulation would temporally precede H₂O₂ accumulation. Instead, our data indicate that H₂O₂ levels peak before O₂ ⁻ levels peak. Separately from its interaction with SOD, O₂ ⁻ also interacts with nitric oxide (NO) to form peroxynitrite (ONOO⁻), which (like H₂O₂) is known to control traditional cell signaling pathways downstream of O₂ ⁻ levels [50, 51].

Therefore, we next hypothesized that instead of driving H₂O₂ formation, O₂ ⁻ reacts with NO to promote ONOO⁻ signaling. If supported, we would predict that 1) injury-induced wound site ONOO⁻ accumulation occurs during planarian regeneration, and 2) the pattern of peak ONOO⁻ levels would align temporally with our observed peak of O₂ ⁻. We used the species-specific fluorescent reporter dye 2′,7′-dichlorodihydrofluorescein diacetate (DCDHF) to visualize ONOO⁻ levels during tissue growth (Figure 7A, bottom panels). Our data show that similar to O₂ ⁻, while at 1 h ONOO⁻ was present at the wound site, there was a significant increase in ONOO⁻ levels by 2 h (Figure 7B).

Together, these data reveal that the accumulation of H₂O₂ and O₂ ⁻ are temporally distinct during stem-cell mediated tissue growth. In addition, our results highlight a previously unappreciated role for H₂O₂-independent ROS signaling mechanisms during this process. These data indicate that both H₂O₂ and O₂ ⁻/ONOO⁻ pathways are activated after injury, suggesting that ROS mechanisms during planarian regeneration are reliant on more than one ROS signaling pathway.

The asexual clonal line of Schmidtea mediterranea (CIW4) was maintained in the dark at 18 C. Planarians were kept in ultrapure Type 1 water with Instant Ocean salts at 0.5 g/L (worm water). Animals were fed every third week with liver paste processed from a whole calf liver (antibiotic and hormone free) obtained from Creekstone Farms (Arkansas City, KS). Liver paste was never frozen or thawed more than once before feedings. Worms 2–5 mm in length were used for all experiments and worms were starved at least 1 week before use. Amputations were done as previously described [92] with a dissecting microscope on a custom-made cooling Peltier plate. Trunk fragments were produced via transverse amputation just anterior and posterior to the pharynx, with cuts made at a 90 degree angle to the sagittal plane for consistency in wounding. All untreated controls were held according to field standards in a biological oxygen demand incubator (VWR) at 18 C in the dark.

Experimentally-controlled static WMF exposures were done with custom-built triaxial Helmholtz coils in a μ-metal enclosure (MagShield box) to block external magnetic fields as previously described [15]. Direct electric current to Helmholtz coils was supplied by DC power sources (Mastech HY3005D-3) and was fed through both x and y axis coils to produce a uniform magnetic field. The MagShield box was kept in a temperature-controlled room (20 C). Animals were placed in either 35 or 60 mm Petri dishes in worm water (or in specific media as described in individual assays) in the center of each Helmholtz coil. Magnetic field exposures were performed in the dark always with one coil set at 45 μT (Earth normal average for the geomagnetic field) separated by a μ-metal partition from the other side, where a second coil was set at indicated experimental field strengths from 0 to 900 μT. Before and at the end of each experiment field strengths were confirmed using either a gauss or mG m (AlphaLab models GM1-HS or MGM). Unless otherwise specified, all planarians were exposed to WMFs within 5 min of amputation and then continuously until scoring and imaging at the indicated time. For Figure 3D: total experimental replicates for blastema growth assays were n ≥ 1, with total biological replicates for each condition as follows: 45 μT, n = 164; 0 μT, n = 19; 100 μT, n = 28; 200 μT, n = 25; 300 μT, n = 18; 400 μT, n = 18; 500 μT, n = 17; 600 μT, n = 16; 700 μT, n = 14; 800 μT, n = 11; 900 μT, n = 18.

General ROS and individual species were visualized by the use of cell-permeant live fluorescent reporter dyes. All images were taken ventrally, and animals were kept in the dark while loading dye. For detecting general ROS levels, the oxidative stress indicator dye, 5-(and-6-)-chloromethyl-2',7′-dichlorodihydrofluorescein diacetate (CM-H₂DCFDA; Molecular Probes C6827; excitation, 470 nm; emission, 525 nm) was used. Intact planarians were pre-exposed to the specified WMFs (see above) for 23 h*, at which time they were amputated to produce trunk fragments. Fragments were placed in 25 μM CM-H₂DCFDA (from 10 mM DMSO stock) and returned to the specified WMF for 1 h, at which time regenerates were rinsed 3X in worm water and imaged. Total experimental replicates were n ≥ 1. Total biological replicates were: 45 μT, n = 189; 0 μT, n = 19; 100 μT, n = 17; 200 μT, n = 24; 300 μT, n = 23; 400 μT, n = 26; 500 μT, n = 12; 600 μT, n = 20; 700 μT, n = 22; 800 μT, n = 21; 900 μT, n = 23. (*Note, we have since determined that WMF pre-exposure is not required to obtain the observed WMF effects; see for example our H₂O₂ and O₂ ⁻ WMF data in Figure 5 and Figure 6D–F)

This protocol was used for the remaining dyes, with the following exceptions: H₂O₂ was detected by soaking newly amputated fragments (with no WMF pre-exposure) in 20 μM peroxy orange 1 (Sigma SML0688; from 1 mM DMSO stock; excitation, 470 nm; emission, 525 nm) plus the specified WMF for 1-h prior to imaging. O₂ ⁻ was detected by soaking fragments (with no WMF pre-exposure) for 1-h prior to imaging in 2 μM orange 1 (Enzo Life Sciences ENZ-51012; from 5 mM dimethylformamide stock; excitation, 550 nm; emission, 620 nm). For “normal” (untreated/unexposed) experiments, peroxy orange 1 and orange 1 dye were used as above, but without concurrent WMF exposure. ONOO⁻ was detected by soaking regenerating fragments in 10 μM 2′,7′-dichlorodihydrofluorescein diacetate (DCDHF) for 1-h prior to imaging (Enzo life sciences ALX-610-022-M050; from 10 mM dimethylformamide stock; excitation 502 nm; emission 523 nm). For 2 h timepoints, fragments were cut and allowed to regenerate for 1 h, at which time animals were soaked in dye for another hour before rinsing and imaging. For all time points, animals were rinsed in ice cold worm water 3X to preserve fluorescence. Total experimental replicates for all were n ≥ 2. Total biological replicates for O₂ ⁻ were: 45 μT at 1 h, n = 42; 200 μT at 1 h, n = 25; 500 μT at 1 h, n = 20; untreated at 1 h n = 9; 45 μT at 2 h, n = 53; 200 μT at 2 h, n = 21; 500 μT at 2 h, n = 20; untreated at 2 h, n = 18. Total biological replicates for H₂O₂ were: 45 μT at 1 h, n = 47; 200 μT at 1 h, n = 19; 500 μT at 1 h, n = 25; untreated at 1 h, n = 12; 45 μT at 2 h, n = 58; 200 μT at 2 h, n = 27; 500 μT at 2 h, n = 26; untreated at 2 h, n = 14. Total biological replicates for ONOO⁻ were: untreated at 1 h, n = 30; untreated at 2 h, n = 12.

Fluorescent in situ hybridization (to observe and quantify mRNA expression) was performed as previously described [93], with the following exceptions: Prehybe and hybe used yeast RNA at 1 mg/ml and probe dilution was 0.5 ng/μl with hybridization for 24 h. S. mediterranea riboprobes to Hsp70, Piwi-1, and Agat-1 were generated as described in our previous paper in (15). The regions/primers used were: for Hsp70, a 552 bp region from 5′-GGT​TTT​TGA​TTT​GGG​TGG​TG to 3′-AGC​TGT​TGC​TAT​GGG​AGC; for Piwi-1, a 2461 bp region from 5′-GAT​CCC​AAT​TTA​AGA​CCA​AGA​AGA​G to 3′-TTT​TTA​TGT​ATT​CGA​TTA​AAA​AAA​A; and for Agat-1, 404 bp from 5′-GGA​GTT​AAA​GTG​TCC​ATC​CAG to 3′-GTT​GCT​AAC​CTG​ACT​GAC​ATG​C. Total experimental replicates for all were n ≥ 1. Total biological replicates for Hsp70 riboprobe were: 45 μT, n = 11; 200 μT, n = 11; 500 μT, n = 11. For Piwi-1 riboprobe: 45 μT (200 μT control), n = 4; 200 μT, n = 4.45 μT (500 μT control), n = 7; 500 μT, n = 7. For Agat-1 riboprobe: 45 μT (200 μT control), n = 5; 200 μT, n = 5; 45 μT (500 μT control), n = 5; 500 μT, n = 5. Labeling of miotic cells by immunostaining was performed as previously described [94], with anti-pH3 (Sigma/Millipore 04-817; 1:25) as the primary antibody. A goat anti-rabbit horseradish peroxidase (Invitrogen 65-6120) with TSA Cyanine 3 (Cy3)-tyramine (PerkinElmer; 1:50) amplification was used as the secondary antibody. All experiments were run once with controls. For Piwi-1 and Agat-1 a second experiment was run for representative photos. Total biological replicates: 45 μT, n = 39; 200 μT, n = 25; 500 μT, n = 19.

ROS production was inhibited with diphenyleneiodonium chloride (DPI; Sigma D2926). Endogenous ROS in the form of H₂O₂ was administered by soaking planarians in 400 μM H₂O₂ (diluted from 30% stock; Sigma 216763). For Figures 6A–C: Intact worms were presoaked in 10 μM DPI (from 1 mM DMSO stock) for 24 h. Animals were amputated to form trunk fragments, then placed in worm water (DPI) or 400 μM H₂O₂ (DPI + H₂O₂) and allowed to regenerate at 18 C. At 3 days post injury, animals were imaged and scored for blastema size. Controls were pre-exposed to an equal amount of DMSO, then placed in worm water after amputation. Experiments were run at least 1 time. Total biological replicates: DMSO controls, n = 18; DPI, n = 8; DPI + H₂O₂, n = 10. For Figures 6D–F: 400 μM H₂O₂ was added after amputation concurrent with 200 μT WMF exposure. Experiments were run twice. Total biological replicates: 45 μT, n = 35; 200 μT, n = 28; 200 μT + H₂O₂, n = 23. For Figures 6G–I, animals were presoaked in 400 μM H₂O₂ (Exogenous H₂O₂) or worm water (Untreated Controls) for 24 h prior to amputation and then returned to H₂O₂ (or worm water for controls) for 1 h, at which time all animals were rinsed 3X in worm water and placed in the O₂ ⁻ dye orange 1 (as described above) for an additional hour prior to imaging at 2 h post amputation. Experiments were run once. Total biological replicates: H₂O₂, n = 10; untreated, n = 12.

A Zeiss V20 Fluorescence Stereomicroscope with an AxioCam MRc or MRm camera and ZEN (lite) software was used for image collection. All live images were taken while regenerates were moving (fully extended) to prevent skewing blastema size/signal intensity due to scrunching. For blastema size, animals were imaged in 100 mm Petri dishes with worm water. For live dyes, animals were imaged in 35 mm FluoroDishes (WPI FD35-100) with 25 mm round no. 1.5 coverslips (WPI 503508). For the general ROS dye CM-H₂DCFDA, heat maps were generated using the standard rainbow lookup table (LUT) to visualize signal intensity. For each assay, samples were imaged at the same magnification and exposure levels to prevent confounding variables during comparisons (i.e., acquisition conditions were kept constant across an experiment between control/treated and/or all different time points). Photoshop (Adobe) was used to orient and scale images (and improve clarity for morphology only). No data was added or subtracted. Original images available by request.

The magnetic lasso tool in Photoshop (Adobe) was used to generate total pixel counts of the anterior blastema (white tissues) and total regenerate (entire worm including blastema). To account for any variation in worm size, blastema was calculated as percent of total body size: (blastema size/body size) x 100. The magnetic lasso tool was also used to measure gray mean values (signal intensity) of fluorescent dyes at the anterior blastema. To account for any variation in dye loading, signal intensity was calculated as the difference between signal at the blastema versus signal from the middle of the regenerate (the pharyngeal region): blastema – pharyngeal region. Cell counts of pH3+ were done using the RTNC plugin tool with ImageJ. Number of mitotic cells was expressed as cells per mm² of the entire regenerate, with total area measured using the magnetic lasso tool (as before). Significance: either two-tailed Student’s t-test with unequal variance (Microsoft Excel or GraphPad Prism 9); or one-way analysis of variance (ANOVA) with Tukey’s multiple comparison test (GraphPad Prism 7).