Elemental and biomolecular changes in serum as indicators of SPION action in the body: analysis using TXRF, FTIR and Raman spectroscopy Authors: Aleksandra Wilk, Karolina W. Lakomy, Zuzanna Setkowicz, Karol Szary, Regina Stachura, Katarzyna M. Marzec, Marzena Rugiel, Joanna Chwiej Published: 2025 ABSTRACT Superparamagnetic iron oxide nanoparticles (SPION) have enormous potential for the use in medical diagnostics and therapy, including theranostics and personalized medicine. However, their clinical implementation remains limited, partly due to insufficient understanding of their systemic biocompatibility, metabolism and long-term physiological impact. In this study, we investigated the molecular and elemental alterations in blood serum following SPION administration, with the goal of identifying sensitive biomarkers of nanoparticle-induced biological responses. PEGylated SPION with a 10 nm magnetite core were intravenously administered to male and female Wistar rats. Serum samples were collected at 2 h, 24 h, and 7 days post-injection. A multimodal analytical approach was employed, combining Fourier-transform infrared (FTIR) and Raman spectroscopy to assess biomolecular changes, with total reflection X-ray fluorescence (TXRF) spectroscopy for quantitative elemental profiling. Spectroscopic analysis revealed significant, time-dependent alterations in the serum biochemical landscape. FTIR data indicated a progressive decrease in the lipid-to-protein ratio, suggesting lipid remodeling. Raman spectra showed increased hemoglobin-associated signals, consistent with increase hemolysis and indicated changes in lipid saturation as well as protein secondary structure. TXRF results revealed marked reductions in serum zinc and selenium levels—key antioxidants—alongside dynamic fluctuations in copper and calcium, often exhibiting sex-dependent trends. These findings suggest that SPION metabolism triggers coordinated physiological responses involving erythrocyte damage, oxidative stress, inflammatory signaling and disruption of elemental homeostasis. The applied multimodal spectroscopic approach demonstrates strong potential for non-invasive, high-resolution monitoring of nanoparticle-induced systemic effects and supports its future utility in nanotoxicology and translational nanomedicine.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-025-29434-z. Introduction Due to their properties, superparamagnetic iron oxide nanoparticles (SPION) have found various applications in numerous branches of engineering and industry1,2. The unique characteristics, such as the nanometer size of SPION, the high surface-to-volume ratio, colloidal stability, and especially their magnetic properties, provide opportunities to modify and use these particles also in medicine both for diagnostic and therapeutic purposes1,3,4. SPION can be manipulated by applying an external magnetic field gradient. When conjugated with therapeutic agents, they can be directed to specific sites within the body, offering a promising strategy for targeted drug delivery4,5. They can also generate heat in response to an alternating magnetic field through the mechanisms such as Brownian and Néel relaxation or the hysteresis loop effect6. This localized heating, known as magnetic hyperthermia, has the potential for application in cancer treatment. By delivering nanoparticles (NP) to the tumor site using a gradient of external magnetic field, magnetic hyperthermia can be utilized to selectively destroy pathological tissues7,8. SPION affect both longitudinal (T1) and transverse (T2) relaxation times, making them effective contrast agents for magnetic resonance imaging (MRI). These novel contrast agents may offer a potential alternative to the commonly used gadolinium-based contrasts causing complications in patients with renal impairment9–11. Although SPION and their potential biomedical applications have been extensively studied over the past few decades, their commercial use in medicine remains limited. This is primarily due to concerns about NP toxicity that arise from their unique metabolic pathways. The metabolism of SPION in living organisms, particularly their final fate and metabolites, is still not fully understood. Various phenomena may be responsible for the potential toxicity of SPION. One of them is a significant local increase in free iron ion concentration leading to elevated levels of reactive oxygen species and in consequence to the oxidative stress. SPION themselves could have a cytotoxic effect on cells, inducing pathological processes by both mechanical damage to cellular components (destruction of the cell membrane, nucleus or mitochondria) and altering the balance of chemical processes occurring in cells12–15. Some new light on the metabolism and potential toxicity of SPION may be shed by the identification of the elemental and biomolecular abnormalities that appear in the serum after SPION administration. Blood circulates throughout the whole body transporting drugs and their metabolites to all tissues and organs. It can be easily and minimally invasively collected allowing for frequent monitoring and tracking of the changes in its parameters over time. The changes in blood composition may be a source of information on drug effectiveness and safety for an individual patient, which is crucial in the context of personalized medicine16–19. To verify the hypothesis that the biomolecular composition of serum may be the source of important information on SPION action, their biokinetics and fate in the body, the serum samples were taken from Wistar rats injected with the PEG-coated SPION containing a 10 nm magnetite core and naive controls. To follow the dynamics of these changes three time points from SPION injection—2 and 24 h and 7 days—were examined. Given the differences in iron metabolism between sexes, the experiment was conducted separately on male and female rats. The aforementioned differences stem from varying iron requirements, the influence of sex hormones on its metabolism20–23, differences in the absorption, storage and removal of iron, and, finally, pathological conditions that affect iron levels and occur with different probabilities depending on sex24. To determine SPION-induced elemental changes occurring in serum total reflection X-ray fluorescence (TXRF) was utilized. Its biochemical anomalies were examined using Raman spectroscopy and Fourier-transform infrared (FTIR) spectroscopy. TXRF is a well-established technique for multielemental analysis of a variety of materials, including biological, medical, and biomaterials. For average and higher-Z elements, it is a reliable technique characterized by good replicability, high precision and low detection limits25–29. On the other hand, the methods of vibrational spectroscopy, such as FTIR and Raman spectroscopy, provide molecular information on the tested samples. Both these techniques suit well for the analysis of biological material providing insights into the content and structure of proteins, lipids, nucleic acids and other compounds in the examined samples30–34. Raman spectroscopy is also a valuable tool for studying inorganic materials, including iron oxides. Our earlier studies showed that it allows us to follow internalization and fate of IONP into cells and biomolecular changes in their composition occurring as a result of NP action30,35,36. Results FTIR and Raman spectroscopy The spectra presented in Fig. 1 were obtained in transflection mode using FTIR spectroscopy performed on rat serum samples deposited on MirrIR low-e reflective slides. These average spectra represent groups of male and female rats administered with SPION at three distinct time points post-injection: 2 h (M2H and F2H), 24 h (M24H and F24H), and 7 days (M7D and F7D), compared to untreated controls (NM and NF). The solid lines indicate the group-averaged spectra, while the shaded regions denote the standard deviation, providing a visual estimate of intra-group spectral variability. The tentative vibrational assignments of IR-active bands and their molecular origins are summarized in Table 1. Notably, the 1651 cm⁻1 band corresponds to amide I vibrations, primarily from serum albumin and globulins, while the lipids fraction includes mostly CH₂ and CH₃ stretching vibrations between 2800 and 3000 cm⁻1.Fig. 1The comparison of average IR spectra of serum samples obtained for male and female rats 2 h, 24 h and 7 days from SPION injection. The solid line stands for an average spectrum for the population whilst the shadow represents standard deviation of the spectra measured for particular individuals of the given group.Table 1The tentative assignment of the bands analyzed in FTIR and Raman spectra to the vibrations of bonds or functional groups of compounds present in the serum samples.FTIR band [cm⁻1]AssignmentMolecular originReference(s)1651Amide I (C=O stretching, α-helix)Serum proteins37,382850CH₂ symmetric stretchingLipids37–402870CH3 symmetric stretchingLipids and proteins37–402925CH₂ asymmetric stretchingLipids37–402955CH3 asymmetric stretchingLipids and proteins37–402800–3000CH₂/CH3 stretchingLipids37–40Raman shift [cm⁻1]AssignmentMolecular origin/spectral type*Reference(s)750ν₁ porphyrin ring breathingHemoglobin/RRS41,421004symmetric ring breathingPhenylalanine (proteins)431130ν₂ skeletal mode (C–C/C–N in porphyrin ring)Hemoglobin/RRS41,421309CH₂ twisting (methylene groups)Saturated lipids/RS44–461339CH₂ bending (methylene scissoring)Lipids, proteins/RS43,441445CH₂ scissoringLipids and proteins/RS471598ν₁₀ C=C stretching (porphyrin ring)Hemoglobin/RRS42,481655Amide I (C=O stretch, α-helix)Serum proteins (albumin, globulins)/RS492880CH₂ asymmetric stretchingLipids, proteins/RS43,442928CH₂ asymmetric stretchingLipids/RS443060= CH stretching (aromatic and olefinic groups)Aromatic amino acids (Tyr, Phe)/RS43*RS, Spontaneous Raman; RRS, Resonance Raman Spectroscopy. The comparison of average IR spectra of serum samples obtained for male and female rats 2 h, 24 h and 7 days from SPION injection. The solid line stands for an average spectrum for the population whilst the shadow represents standard deviation of the spectra measured for particular individuals of the given group. Table 1The tentative assignment of the bands analyzed in FTIR and Raman spectra to the vibrations of bonds or functional groups of compounds present in the serum samples.FTIR band [cm⁻1]AssignmentMolecular originReference(s)1651Amide I (C=O stretching, α-helix)Serum proteins37,382850CH₂ symmetric stretchingLipids37–402870CH3 symmetric stretchingLipids and proteins37–402925CH₂ asymmetric stretchingLipids37–402955CH3 asymmetric stretchingLipids and proteins37–402800–3000CH₂/CH3 stretchingLipids37–40Raman shift [cm⁻1]AssignmentMolecular origin/spectral type*Reference(s)750ν₁ porphyrin ring breathingHemoglobin/RRS41,421004symmetric ring breathingPhenylalanine (proteins)431130ν₂ skeletal mode (C–C/C–N in porphyrin ring)Hemoglobin/RRS41,421309CH₂ twisting (methylene groups)Saturated lipids/RS44–461339CH₂ bending (methylene scissoring)Lipids, proteins/RS43,441445CH₂ scissoringLipids and proteins/RS471598ν₁₀ C=C stretching (porphyrin ring)Hemoglobin/RRS42,481655Amide I (C=O stretch, α-helix)Serum proteins (albumin, globulins)/RS492880CH₂ asymmetric stretchingLipids, proteins/RS43,442928CH₂ asymmetric stretchingLipids/RS443060= CH stretching (aromatic and olefinic groups)Aromatic amino acids (Tyr, Phe)/RS43*RS, Spontaneous Raman; RRS, Resonance Raman Spectroscopy. The tentative assignment of the bands analyzed in FTIR and Raman spectra to the vibrations of bonds or functional groups of compounds present in the serum samples. *RS, Spontaneous Raman; RRS, Resonance Raman Spectroscopy. The comparison of the FTIR spectra reveals a consistent decrease in the lipid-to-protein ratio following SPION administration, evident through the reduction in integrated absorbance within the 2800–3000 cm⁻1 region relative to the amide I band (Fig. 3). This trend appears progressive and suggests time-dependent metabolic or structural changes in circulating lipids and/or proteins. This finding may reflect either SPION-induced oxidative degradation of serum lipids or altered lipid metabolism, potentially linked to systemic responses such as inflammation or hepatic processing. On the other hand after SPION injection opsonization process manifesting by increased proteins level in serum may also occur. Raman spectroscopy was performed on the same plasma samples using a 532 nm laser excitation source. Although MirrIR substrates introduced additional background signals—predominantly due to substrate-derived vibrational modes (~ 670, ~ 570, ~ 450, ~ 328, and ~ 280 cm⁻1)—spectral contributions from the plasma were still clearly distinguishable (see Figure S1 of the Supplementary materials). These bands are likely associated with the vibrational modes of Si–O–Si bending (~ 670 and ~ 450 cm⁻1) and metal–oxygen vibrations (~ 570, ~ 328, ~ 280 cm⁻1)50. A broad band near 1100 cm⁻1 is associated with stretching modes of SiO4 tetrahedra due to silicon motions in the substrate overlapped with the biologically relevant 1130 cm⁻1 signal50. Therefore, Raman spectra of plasma are presented starting from 800 cm⁻1, and the ~ 1130 cm⁻1 region was excluded from the analysis. The spectra shown in Fig. 2 correspond to average Raman spectra of blood plasma measured across all individuals within each rat population. The tentative vibrational assignments for these bands originated from blood plasma are provided in Table 1. Raman spectra of rat blood plasma acquired under 532 nm laser excitation revealed distinct molecular signatures of spontaneous Raman scattering (RS) features originating from blood plasma constituents. These included characteristic vibrational bands consistent with the known biochemical profile of plasma proteins and lipids. In addition, resonance-enhanced Raman bands attributable to trace hemoglobin contamination were detected, likely due to trace amounts of hemoglobin introduced by erythrocyte hemolysis before serum preparation42.Fig. 2Comparison of average Raman spectra of serum samples from male and female rats at 2 h (M2H and F2H), 24 h (M24H and F24H), and 7 days (M7D and F7D) after SPION injection, relative to untreated controls (NM and NF). Solid lines represent the mean Raman spectra for each group, while shaded areas indicate the standard deviation across individual samples. The average Raman spectra presented correspond to the eight experimental groups, each comprising six animals (N = 6). For better visualization of the fingerprint region, the intensities in the 700–1800 cm−1 range were multiplied by a factor of five relative to the CH stretching region. Comparison of average Raman spectra of serum samples from male and female rats at 2 h (M2H and F2H), 24 h (M24H and F24H), and 7 days (M7D and F7D) after SPION injection, relative to untreated controls (NM and NF). Solid lines represent the mean Raman spectra for each group, while shaded areas indicate the standard deviation across individual samples. The average Raman spectra presented correspond to the eight experimental groups, each comprising six animals (N = 6). For better visualization of the fingerprint region, the intensities in the 700–1800 cm−1 range were multiplied by a factor of five relative to the CH stretching region. Raman spectroscopic monitoring of rat plasma collected at 2 h, 24 h, and 7 days post-SPION injection revealed distinct, time-dependent changes in the molecular composition of the serum. Interestingly, the maximum time from SPION-addition was inversely correlated with the intensity of bands attributed to hemoglobin at 750 and 1598 cm⁻1, which are known to be resonance-enhanced under 532 nm excitation. A progressive increase in the intensity ratio of the 1598 cm⁻1 band (ν₁₀ C=C stretching of the porphyrin ring) to the 1004 cm⁻1 band (phenylalanine) was observed across time points (Fig. 3). Although this enhancement occurs under resonance Raman conditions and is therefore not directly quantitative, it strongly suggests an increasing presence of hemoglobin-derived species, likely due to cumulative hemolysis occurring over time following nanoparticle exposure. Although direct UV–Vis spectrophotometric validation was not possible due to serum volume limitations, the consistent increase of hemoglobin-associated Raman bands (750, 1130, 1598 cm−1) strongly supports the interpretation of hemolysis. Future studies will benefit from complementary UV–Vis measurements to further validate these findings.Fig. 3The ratios of integral intensities for selected FTIR (a) and Raman (b–d) bands were analyzed to assess the alterations in serum samples taken from male and female rats at 2 h (M2H and F2H), 24 h (M24H and F24H), and 7 days (M7D and F7D) after SPION injection, relative to untreated controls (NM and NF). The evaluations included band ratios of: (a) 2800–3000/1651, (b) 1598/1004, (c) 1309/1004, (d) 1655/1445 reflecting, respectively, the following parameters of blood plasma: (a) protein-to-lipids ratio, (b) hemoglobin level, (c) lipid saturation or lipid packing, (d) conformational changes in serum proteins or shifts in the protein-lipid balance. Normality was assessed using the Shapiro–Wilk test. Data are presented as box plots (mean, min–max whiskers, standard deviation box), and significance was calculated using Student’s t-test (# for 0.05 < p-value < 0.1, * for p-value < 0.05). The ratios of integral intensities for selected FTIR (a) and Raman (b–d) bands were analyzed to assess the alterations in serum samples taken from male and female rats at 2 h (M2H and F2H), 24 h (M24H and F24H), and 7 days (M7D and F7D) after SPION injection, relative to untreated controls (NM and NF). The evaluations included band ratios of: (a) 2800–3000/1651, (b) 1598/1004, (c) 1309/1004, (d) 1655/1445 reflecting, respectively, the following parameters of blood plasma: (a) protein-to-lipids ratio, (b) hemoglobin level, (c) lipid saturation or lipid packing, (d) conformational changes in serum proteins or shifts in the protein-lipid balance. Normality was assessed using the Shapiro–Wilk test. Data are presented as box plots (mean, min–max whiskers, standard deviation box), and significance was calculated using Student’s t-test (# for 0.05 < p-value < 0.1, * for p-value < 0.05). It was previously reported, that the band at 1445 cm−1 arises from CH2 and CH3 vibrations of both lipids and proteins, while the 1664 cm−1 band corresponds to amide I (C=O) and C=C stretching in lipids. The 1657/1445 cm−1 ratio is therefore a reliable marker of lipid unsaturation only in lipid-rich samples, where protein contribution to the amide I band is negligible51. In complex biological matrices such as serum, this ratio more accurately reflects the relative protein-lipid balance45. In addition, previous reports have identified characteristic bands associated with lipid saturation (1304 cm−1, 1446 cm−1, 2855 cm−1) and lipid unsaturation (1268 cm−1, 1659 cm − 1, 3011 cm−1)46. On this basis, we focused on the 1309/1004 cm−1 ratio, which reflects the CH₂ twisting vibrations of saturated lipids relative to the phenylalanine band. A time-dependent increase in the intensity of the Raman band at 1309 cm−1, corresponding to CH₂ twisting and wagging modes in saturated lipid chains, was observed. This change, alongside other non-resonant vibrational markers, provides insight into the evolving lipid and protein composition of serum following SPION exposure. In particular, the rising 1309/1004 cm−1 intensity ratio suggests an elevated content of saturated lipids or alterations in lipid organization and membrane packing, potentially reflecting systemic responses to nanoparticle-induced stress44,47. Additionally, the progressive increase in the 1655/1445 cm−1 ratio, which compares mainly the amide I band (C=O stretching of α-helical proteins) to CH₂ scissoring vibrations, suggests shifts in the protein–lipid balance and may indicate structural modifications in serum proteins, such as unfolding or changes in their secondary structure32,44,49,51. Although these lipid- and protein-related spectral trends did not reach statistical significance, their consistent directional shift across time points highlights a biologically relevant tendency toward serum matrix remodeling in response to SPION administration. TXRF The results of elemental analysis of serum samples performed with the TXRF are presented in the Fig. 4 whilst the obtained validation parameters of the method are shown in the Table S1 of the Supplementary materials.Fig. 4Box-whisker plots showing the scatter of the K, Ca, Fe, Cu, Zn and Se concentrations in serum for examined rat populations. The frame marks interquartile span, the whiskers stand for minimal and maximal value, whilst the internal line the median concentration for the particular animal group. Statistically significant differences (Mann–Whitney U test, 95% confidence level) comparing to the naive norm were signed with the asterisks. Box-whisker plots showing the scatter of the K, Ca, Fe, Cu, Zn and Se concentrations in serum for examined rat populations. The frame marks interquartile span, the whiskers stand for minimal and maximal value, whilst the internal line the median concentration for the particular animal group. Statistically significant differences (Mann–Whitney U test, 95% confidence level) comparing to the naive norm were signed with the asterisks. The analysis of the data presented in the Table S1 of the Supplementary materials showed that serum concentrations of all examined elements (K, Ca, Fe, Cu, Zn and Se) were significantly higher that the calculated limits of detections and therefore might be the subject of further comparative study. The median K and Ca concentrations in the serum samples taken from animals subjected to SPION were lower than control ones. However, the observed relations were statistically relevant only for Ca and in case of male individuals. Surprisingly, the Fe level in the serum did not show statistically relevant differences between exposed and normal animals, with the exemption of female rats on the 7th day after treatment. However, some trends, different for male and female rats, were noticed here. In males serum concentration of the element tended to increase, whilst in females to decrease as a result of the exposure. Cu concentration in serum presented similar pattern of changes in males and females. Two hours from the injection of SPION, the level of the element was significantly lower in the treated rats. It was increasing between 2nd and 24th hour after exposure, and then it decreased again in the 3rd observation period. 7 days from NP injection the serum concentration of Cu in exposed rats of both sexes was significantly lower than that measured for the appropriate control group. Although the progress of Cu changes in serum was similar in both sexes, the concentration of the element was usually around 20% higher in female than male rats. The opposite situation was found for Zn which content in serum was systematically higher in males. What is more, Zn and also Se showed statistically relevant decrease of concentration in case of all groups of exposed rats. Discussion This study evaluated the biochemical and elemental alterations in rat blood serum induced by the intravenous administration of PEG-coated SPION. To achieve this, FTIR and Raman spectroscopy, and TXRF were applied to serum samples collected from male and female Wistar rats at 2 h, 24 h and 7 days post-injection. Untreated animals served as controls. The applied methods enabled the time-resolved monitoring of molecular and elemental changes, with particular attention to sex-specific differences. The combined spectroscopic and elemental analyses indicate that SPION exposure triggers a complex cascade of systemic responses reflected in the molecular and elemental composition of rat serum. These responses appear to involve immune activation, oxidative stress, erythrocyte damage and metal homeostasis dysregulation, many of which are known to be mediated by cytokines such as interleukin-6 (IL-6)52,53. Combined FTIR and Raman spectroscopic analyses reveal that SPION exposure induces time-dependent biochemical changes in rat serum, reflecting both metabolic adaptation and systemic stress responses. FTIR data indicate a progressive decrease in the lipid-to-protein ratio, which may result from lipid consumption/redistribution driven by oxidative or inflammatory processes, or from the opsonization process. Complementary Raman analysis shows a time-related increase in hemoglobin-associated bands, consistent with progressive erythrocyte hemolysis linked to SPION degradation or clearance. An elevated 1309/1004 cm⁻1 ratio points to an increase in saturated lipid content or changes in lipid packing, while a rise in the 1655/1445 cm⁻1 ratio may reflect conformational changes in serum proteins or shifts in the protein-lipid balance. Although the last two trends did not reach statistical significance, they suggest dynamic remodeling of the serum molecular profile over time. Altogether, the data support the notion that SPION metabolism affects erythrocyte stability, lipid composition and protein structure in an interrelated manner. These findings highlight the value of combining IR and Raman spectroscopy to monitor subtle, yet biologically meaningful, nanoparticle-induced effects in vivo. The observed biochemical changes in serum following SPION administration may result from several interconnected mechanisms that collectively modify its lipid-protein composition. Iron oxide nanoparticles and the iron ions they release can lead to increased production of reactive oxygen species (ROS), which in turn induce lipid peroxidation and oxidation of amino acids in plasma proteins. This results in chemical modifications of proteins, such as carbonylation and glycation, altering their physicochemical and functional properties54. Disturbances in the protein-to-lipid ratio in serum may also arise from the formation of a so-called protein corona upon introduction of SPION into the bloodstream. Coating nanoparticles with plasma proteins, including opsonins that mark them as “foreign bodies” and facilitate their recognition and clearance by immune cells, may increase the relative proportion of proteins compared to lipids in plasma55–57. Furthermore, changes in the protein-to-lipid ratio, fatty acid saturation levels and lipid packing density may also originate from systemic inflammatory responses or redistribution of lipids by the spleen and liver after nanoparticle uptake by the reticuloendothelial system55,58. TXRF analysis revealed significant disturbances in trace element concentrations, providing further insight into the systemic effects of SPION. Ca levels significantly decreased in male rats, with no consistent changes in females. Ca plays a vital role in immune signaling, particularly in lymphocyte activation and may be modulated by IL-6–stimulated parathyroid hormone activity, leading to reduced serum Ca during systemic inflammation52,53,59,60. Zn levels declined significantly over time in both sexes. Zn plays a critical role in antioxidant defense and immune function and its depletion likely reflects inflammatory redistribution, mediated by IL-6-driven hepatic Zn sequestration and metallothionein expression61–63. Se concentrations also dropped significantly in all treated groups and remained low throughout the observation period. As a key component of selenoproteins (e.g. glutathione peroxidase), Se depletion may reflect increased antioxidant demand or redistribution to stressed tissues, as suggested in inflammatory models64,65. Cu levels followed a biphasic pattern: decreased at 2 h post-injection, increased at 24 h, and then dropped again by day 7. These fluctuations likely mirror ceruloplasmin regulation, an acute-phase Cu-binding protein modulated by liver function and inflammatory cytokines such as IL-666–68. The enzyme’s role in Fe metabolism further ties Cu dynamics to the observed changes in hemoglobin and Fe. Fe levels increased slightly in males but not significantly, and in females, levels remained unchanged. Given that serum Fe is predominantly derived from hemolysis, these observations support the Raman findings and reinforce the conclusion that serum Fe is not a reliable direct indicator of SPION biodistribution, as previously reported69,70. Overall, the co-evolution of molecular and elemental biomarkers following SPION administration reveals a consistent and time-dependent physiological response involving several interconnected processes. The progressive enhancement of hemoglobin-associated Raman bands indicates erythrocyte damage and hemolysis, likely linked to SPION-induced membrane destabilization or oxidative stress. Simultaneously, changes in lipid-related spectral features observed in both FTIR and Raman spectra reflect altered lipid metabolism and increased serum lipid saturation. Modifications in protein structure are suggested by shifts in amide I and CH₂ band ratios, pointing to conformational rearrangements or changes in the protein-lipid balance. Elemental analysis further supports this picture, with evidence of trace element redistribution—particularly the depletion of Zn and Se—and Cu fluctuations consistent with inflammatory responses. These alterations appear to result from the complex interplay between nanoparticle clearance, hepatic processing, immune activation and oxidative pressure. Notably, many of these changes exhibit sex-specific patterns: male rats showed greater susceptibility to disturbances in Ca and Fe homeostasis, whereas females demonstrated distinct variations in Cu and Zn regulation. These findings underscore the importance of considering sex as a biological variable in the toxicological assessment of nanomaterials and highlight the value of integrated spectroscopic approaches for capturing the multifaceted impact of SPION on systemic biochemistry. A potential limitation of this study is the relatively small number of animals used, which was dictated by ethical considerations aimed at minimizing animal use. Nevertheless, all examined animals were of the same age, had similar body masses, were maintained under identical living conditions, and underwent the same procedures within each group. Therefore, the observed variability in elemental concentrations and biochemical parameters within populations can be attributed primarily to natural biological variation rather than experimental inconsistencies. What is more, our findings demonstrate clear and consistent trends, as evidenced by the agreement across three independent analytical techniques (FTIR, Raman, TXRF), which substantially reduces the likelihood of random bias influencing the results. Materials and methods Animals The experiment was conducted and reported in accordance with the ARRIVE guidelines for animal research (https://arriveguidelines.org). The husbandry of Wistar rats, as well as all the procedures using rodents were performed in the Laboratory of Experimental Neuropathology at the Institute of Zoology and Biomedical Research of the Jagiellonian University (Krakow, Poland). The experiment was approved by the 2nd Local Institutional Animal Care and Use Committee in Krakow (agreement no. 318/2020). All methods were performed in accordance with the relevant guidelines and regulations. Throughout their lives, the animals stayed in the conditions of controlled temperature (20 ± 2 °C) and lighting (12-h light and dark cycle). They had also the unlimited access to food (Labofeed H Standard, Morawski) and water. The rats of both sexes were included in the study. Each of 8 animal groups consisted of 6 individuals. These were 4 groups of males and 4 groups of females (M and F groups) that differed with the time passing from the SPION injection to the blood collection. The blood was taken 2 and 24 h and 7 days post-injection (M2H, M24H, M7D and F2H, F24H, F7D groups in case of males and females, respectively). Besides animals subjected to the SPION action, naive control rats of both sexes were examine (NM and NF groups). Median body weights for experimental and control groups, provided in Table 2, were consistent with physiological norms for Wistar rats at corresponding developmental stages.Table 2Median body masses of rats from experimental and control groups with corresponding interquartile spreads.GroupNFF2HF24HF7DMedian mass [g]182.5167.2159.8169.8Interqartile spread [g]6.410.45.88.5GroupNMM2HM24HM7DMedian mass [g]252.4264.7295.7267.2Interqartile spread [g]17.045.333.49.5 Table 2Median body masses of rats from experimental and control groups with corresponding interquartile spreads.GroupNFF2HF24HF7DMedian mass [g]182.5167.2159.8169.8Interqartile spread [g]6.410.45.88.5GroupNMM2HM24HM7DMedian mass [g]252.4264.7295.7267.2Interqartile spread [g]17.045.333.49.5 Median body masses of rats from experimental and control groups with corresponding interquartile spreads. Nanoparticles solution injection PEG-coated SPION with the 10 nm magnetite core (Sigma-Aldrich, Merck) were dispersed in Ca and Mg ions-free PBS (Corning®) a day prior to further procedures. The concentration of Fe in the dispersion was 0.5 mg Fe/mL and the volume of the solution administered intravenously through the tail vein to rat at the 60th day of its postnatal life was equal to 0.25 mL. During the injection, the animals were anesthetized with the isoflurane (Aerrane, Baxter Poland) which dose was adjusted to maintain an analgesic effect. After the injection of SPION, the administration of isoflurane was ceased, and the wound was pressured until the bleeding stopped. Sample collection and preparation After appropriate period of time passing from NP injection the rats were sacrificed by administering Morbital (Morbital 133.3 mg/mL, Biowet) at doses corresponding to their weights. For each animal a sample of blood was taken recta ex animo in the amount allowed by the physiology (usually from 1 to 2 mL). Blood samples were centrifuged for the serum separation. The serum drawn from above the haematocrit was placed in a new test tube, frozen and kept at the temperature not higher than − 20 °C till the measurements or further procedures. For FTIR and Raman measurements, the serum samples were thawed, vortexed and smears were made on the CaF₂ windows and MirrIR silver-coated low-e microscope slides (Kevley Technologies, USA) and then dried. Quantitative elemental analysis using TXRF was based on the internal standard method. For this purpose, 50 µL of 100 ppm gallium solution was added to 300 µL of serum and mixed thoroughly. For each sample 2 replicates were prepared by putting 2.5 μL of the obtained solution on individual quartz glass carrier. Such prepared samples were drying on a heating plate and measured in the TXRF spectrometer. Apparatus Both FTIR and Raman measurements were carried out at the Laboratory of Atomic and Molecular Biospectroscopy in the Faculty of Physics and Applied Computer Science (AGH University of Krakow, Poland). FTIR measurements were carried out in transflection mode using ThermoScientific Nicolet iN10 MX infrared microscope. The system is equipped with ceramic IR source, Ge/KBr beamsplitter and 3 detectors, including nitrogen-cooled MCT detector which was used in this study. The size of the IR beam reduced with the use of square aperture was 150 × 150 µm2. Each spectrum was an average of 32 individual scans recorded for the wavenumber range of 4000–900 cm−1 with the spectral resolution of 8 cm−1. The acquisition of IR absorption spectra, as well as preliminary spectral analysis, were performed with OMNIC Picta software (version 8.1). For each serum sample 50 spectra were recorded, subjected to preprocessing including: atmospheric correction, base line correction via asymmetric least squares and vector normalisation, and then averaged to obtain the absorption spectrum characteristic for given individual. Raman measurements were performed with a confocal WITec Alpha300R Raman microscope (WITec GmbH, Ulm, Germany) equipped with the UHTS 300 spectrometer (600 gratings/mm), an air-cooled solid-state laser operating at wavelength of 532 nm, thermoelectrically cooled spectroscopy camera and Zeiss EC Epiplan-Neofluar DIC objectives. The laser power was set to 10 mW. Spectrum integration time was equal to 0.5 s. Raman measurements and initial preprocessing of the obtained spectra were done using WITec software (WITec Project Plus 5.3). Further processing, including background subtraction, normalisation, and statistical evaluations were carried out using OriginPro software (version 9.7.5.184). For each serum sample 50 Raman spectra were recorded, subjected to processing including cosmic rays reduction and base line correction via asymmetric least squares, and then averaged to obtain the spectrum characteristic for given individual. TXRF measurements were performed at the Laboratory of X-ray Methods of the Centre for Research and Analysis at the Jan Kochanowski University in Kielce. S2 PICOFOX TXRF spectrometer (Bruker Nano) equipped with the Mo-anode X-ray tube working at the voltage of 50 kV and current 0.4 mA was used for elemental analysis of serum samples. The acquisition time of spectrum was 1000 s. The measured spectra were analysed using Picofox Spectra 7 (Bruker Nano) software. The concentrations of examined elements (K, Ca, Fe, Cu, Zn and Se) in serum samples were determined using the internal standard method what was earlier described in details in the other our papers28,71–73. The limit of the detection \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${LOD}_{xi}$$\end{document} of the element \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$x$$\end{document} in serum sample \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$i$$\end{document} was calculated via Eq. (1):1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${LOD}_{xi}=\frac{3 {C}_{xi} \sqrt{{N}_{BGxi}}}{{N}_{xi}}$$\end{document}where: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${C}_{xi}-$$\end{document} concentration of the element \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$x$$\end{document} in the sample \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$i$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\left[\mu g/g\right]$$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${N}_{BGxi}-$$\end{document} background area under the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$K-\alpha$$\end{document} line of the element \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$x$$\end{document} in the spectrum of sample \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$i$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$[cts]$$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${N}_{xi}-$$\end{document} net area of the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$K-\alpha$$\end{document} line of the element \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$x$$\end{document} in the spectrum of sample \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$i$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$[cts]$$\end{document}. To estimate the final value of the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${LOD}_{x}-$$\end{document} limit of detection for the element \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$x$$\end{document}, all the obtained \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${LOD}_{xi}$$\end{document} values were averaged. The intra-day precision of the TXRF measurements was estimated for each element as a coefficient of variation. This was done independently for male and female rats, according to a following Eq. (2):2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$CV=\frac{1}{{x}_{i}}\sqrt{\frac{{\sum }_{1}^{n}{({x}_{ij}-{x}_{i})}^{2}}{n (n-1)}} 100\%$$\end{document}where: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$j -$$\end{document} measurement number, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${x}_{ij}-$$\end{document} the concentration of the element \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$i$$\end{document} in a measurement \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$j$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\left[\mu g/g\right]$$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${x}_{i}-$$\end{document} the mean value of the concentration of the element \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$i$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\left[\mu g/g\right]$$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$n -$$\end{document} the overall number of measurement. Statistical analysis In the case of biochemical analyses, for each serum sample 50 IR spectra and 50 Raman spectra were recorded, resulting in a total of 300 spectra per experimental group (6 animals × 50 spectra) for each technique. After preprocessing, spectral integrations were performed and the intensity ratios of the relevant bands were calculated. The Shapiro–Wilk test (n = 300) confirmed a normal distribution of all calculated parameters. Therefore, Student’s t-test was applied for comparisons between groups, using the mean as the expected value for the population. For elemental analyses, serum concentrations were determined once for each animal, yielding six values per group. Due to the limited sample size, reliable testing of distribution normality was not possible, and parametric tests could not be applied. Instead, the Mann–Whitney U test was used to assess differences between group medians, with significance levels set at p < 0.05 and p < 0.1. This test represents the most sensitive nonparametric alternative to the t-test and, under certain conditions, may even provide greater statistical power to reject the null hypothesis. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1. Supplementary Material 1.