Health Risk Assessment and Accumulation of Potentially Toxic Elements in Capsella bursa-pastoris (L.) Medik

Abstract

Capsella bursa-pastoris (L.) Medik (C. bursa-pastoris) is an underexplored medicinal herb and bioindicator of potentially toxic elements (PTEs). Its broad traditional utilization combined with its high capacity for PTE accumulation may endanger human health. Herein, we investigated the concentrations and mobility of PTEs (Ba, Co, Cr, Cu, Fe, Mn, Ni, Sr, and Zn) in the urban soil–C. bursa-pastoris system and comprehensively assessed potential health risks associated with exposure to contaminated soils, plant and herbal extracts. Cu, Zn, Sr, and Mn were the most abundant in soils and predominantly phytoavailable. The calculated values of the geo-accumulation index (I_geo) indicated moderate to heavy Cu, Zn, and Sr contamination in the soil. C. bursa-pastoris demonstrated two strategies for PTEs—the exclusion of Ba, Cr, Mn, and Sr, and the accumulation of Cu, Ni, Co, and Fe. Principal Component Analysis (PCA) classified samples from four cities based on the PTE levels in soils, plants, and herbal extracts. Although plant tissues contained elevated levels of PTEs, the estimated daily intake (EDI), target hazard quotient (THQ), and lifetime carcinogenic risk (LCR) demonstrated no significant health risks from consuming C. bursa-pastoris and its extracts. The obtained results indicated the higher sensitivity of children to the hazardous effects of PTEs compared to adults. Extensive risk assessments of polluted soils and inhabiting plants are crucial in PTE monitoring. This study underscored its importance and delivered new insights into the contamination of medicinal herbs, aiming to contribute to implementing safety policies in public health protection.

1. Introduction

Medicinal plants have been traditionally used worldwide for thousands of years as folk remedies for treating various diseases. According to World Health Organization (WHO) reports, almost 80% of the world population consumes herbal medicines as primary healthcare or supplements combined with other medications [1]. Although considered safer for human nutrition and more natural than synthetic medicines, medicinal herbs are chemically complex and may absorb high amounts of toxic elements [2,3]. Ensuring the safety and maintaining the therapeutical effects of medicinal herbs remain challenges.

Safety concerns arose from the evidence of toxic elements in edible and medicinal plants, whereby increasing environmental pollution plays a key role [4,5]. Quality control requirements warranting that concentrations of these elements are within recommended safety limits are regulated and obligatory for herbal product manufacturers [6]. However, the full plant’s elemental composition remains to be considered, although even elements with crucial physiological functions in humans may be toxic in excessive amounts [2]. There is a general, misleading public perception that those elements are safe and without risk, even in cases of oversupply. Many of them have demonstrated a dual nature with a narrow gap between essentiality and toxicity and are thus considered potentially toxic elements (PTEs) [6,7]. PTEs refer to elements that, at certain concentrations, can pose risks to human health, ecosystems, or the environment [8]. Overloads of Cu, Fe, Zn, and Mn are associated with a variety of adverse health effects, such as Parkinson’s disease and hepatocellular cancer and immunosuppressive and neurotoxic impacts [9]. Despite its documented essentiality [10], certain chemical compounds of Cr are considered a pervasive environmental contaminant triggering cancer development and are thus classified as a group 1 carcinogen [11]. Similarly, certain compounds of Ni have been identified as immunotoxin and carcinogenic agents, and its essentiality is still in doubt [12]. Co is essential to humans as a constituent of cobalamin (vitamin B12), while several forms are reported as genotoxic and involved in lung cancer development [13]. Additionally, Sr and Ba are mainly accumulated in the skeleton and may cause the defective mineralization of bones [14,15]. Some widely available supplements may contain even 100-fold higher levels of the recommended daily intakes of these elements [16]. A risk assessment of the human health hazards generated from exposure to PTEs became imperative.

Numerous plant species are known as PTE (hyper-) accumulators [17]. For instance, Noccaea caerulescens and Arabidopsis halleri are hyperaccumulators of Zn, Cd, Pb and Cu [7]. Some of those species are even of medicinal interest. Capsella bursa-pastoris L. Medik, from the Brassicaceae family, is an effective accumulator of toxic elements in polluted environments and represents a promising target for environmental pollution biomonitoring [18]. The Brassicaceae is a family abundant in metal-accumulating crops and medicinal species growing in anthropogenically disturbed and highly polluted sites, whose strong tolerance to PTEs represents a potential health risk [19]. Owing to its complex composition, C. bursa-pastoris, also called shepherd’s purse, provides a broad spectrum of human health benefits such as anticancer, anti-inflammatory, cardioprotective, and antioxidant effects [20]. It is consumed as a tea, applied to the skin as a tincture, added to meals as a spice, or eaten raw [18]. The roots of C. bursa-pastoris may even be used as a ginger substitute [21].

Despite its known tendency to accumulate PTEs, the influence of this accumulation on human well-being is still unknown. To date, no human health risk assessment concerning shepherd’s purse and its preparations has been documented. This study aimed to address this critical research gap, provide an extensive understanding of PTE migration within the soil–C. bursa-pastoris system in highly contaminated urban environments, and deliver critical insights to support the development of environmental technologies and public health policies aimed at mitigating risks associated with contaminated medicinal herbs and their derived products. To the best of our knowledge, such an integrative approach is novel for C. bursa-pastoris.

The specific objectives were to (1) determine the concentration and phytoavailability of selected PTEs (Ba, Co, Cr, Cu, Fe, Mn, Ni, Sr, and Zn) in soils, (2) determine PTE concentrations in both the roots and shoots of C. bursa-pastoris, (3) evaluate the translocation rates of PTEs from soil to plant tissues, (4) analyze PTE content in herbal extracts prepared from the plant’s shoots, (5) explore relationships among PTE levels in soils, plants and extracts, and (6) conduct a comprehensive human health risk assessment related to PTE exposure through contaminated soils, plant tissues, and herbal extracts.

2. Materials and Methods

2.1. Sampling Site Description

Since C. bursa-pastoris is widely distributed in anthropogenically modified areas [22], four differently populated cities in Serbia were selected for sampling—Bor (BO) (44°04′25″ N, 22°05′26″ E) and Vršac (VR) (45°07′00″ N, 21°18′08″ E), chosen to reflect the industrial influence caused by mining activities; Belgrade (BG) (44°49′14” N, 20°27′44” E), under heavy traffic pollution; and Sremska Mitrovica (SM) (44°58′20″N, 19°36′33″E), located close to the nature reserve, assumed to be the least polluted zone (Figure 1). Due to variations in the population density and dominant pollution sources, different levels of anthropogenic influence were expected among the sites. Variations within an individual site could also be anticipated. Thus, each site encompassed three sublocations: 1—city center, 2—midpoint, and 3—periphery.

Bor is the European environmental hotspot; it is considered one of the most polluted cities in Serbia, owing to its comprehensive mining activities. Samples were taken in the residential zone—close to the mining facility (BO1) and 8 km to the south (BO2). A suburb in the northern part was the third sublocati5on. Vršac, a medium-sized city in the northeast of Serbia, was the third site with the following sampling points: VR1—center; VR2—periphery of the city, near the road and the pharmaceutical production complex; and VR3—the foot of the Vršac Mountains, 14 km to the east. Belgrade, the capital and the most populated city, is located in central Serbia. Grassland exposed to heavy traffic located behind the highway close to the city center was chosen as the first and the most polluted point (BG1). The second one was the park near the Danube Quay promenade, about 5 km from the center, while the third sublocation was placed around 7 km to the south in the urban forest. Sample collection in Sremska Mitrovica, the town in the northwest of the country near the nature reserve Zasavica, was carried out in the park in the center (SM1), in the residential zone 1 km to the south, close to the Sava Quay (SM2), and in the suburban area, close to the protected nature reserve Zasavica (SM3).

2.2. Samples Preparation

2.2.1. Soils

The rhizosphere soil of C. bursa-pastoris (up to 20 cm in depth) was sampled for the analysis. Soils were collected in paper bags, transferred to the laboratory, and sieved through a stainless-steel sieve with 2 mm mesh size, according to the standard procedure ISO 11464:2006 (Soil quality—pretreatment of samples for physico-chemical analysis). Soil fraction <2 mm was used for soil analysis. Laboratory samples prepared by the zig-zag method were labeled based on the corresponding sampling site as follows: BG (1–3), BO (1–3), VR (1–3), and SM (1–3).

Pseudo-total content of PTEs in soil was determined by digestion in aqua regia following the USEPA method 3050 [23]. The digestion used a 1:3 ratio of HCl to HNO₃. According to McGrath (1996), a single-step extraction with 0.05 M EDTA [24] was used to assess the phytoavailable fraction of Ba, Co, Cr, Cu, Fe, Mn, Ni, Sr, and Zn.

2.2.2. Plant Material

Approximately 30 specimens of C. bursa-pastoris were collected per sublocation to form the composite sample. After thorough cleaning with tap and distilled water, plants were air-dried at room temperature for two weeks, divided into roots and shoots, i.e., underground and aboveground parts, grounded using agate mortar and pestle, and subjected to further analysis. Microwave digestion of the plant material was performed with 30% H₂O₂ and concentrated HNO₃ (1:7 ratio) using a microwave oven (Ethos 1, Advanced Microwave Digestion System; Milestone, Sorisole, Italy). Digestion parameters were adjusted as follows: warming up to 200 °C for 10 min and heating for 15 min at a constant temperature of 200 °C afterward [25].

2.2.3. Herbal Extracts

Herbal infusions and ethanol tinctures were prepared as described in previous studies [26]. To prepare a water infusion, i.e., herbal tea, 2 g of ground dry plant material (aboveground parts) was covered with 20 mL of boiling distilled water. After 2 min of boiling, the glass was removed from the heater, covered to sit for 30 min, and filtered through a quantitative filter paper (Whatman No. 42). Solutions were transferred into a 25 mL volumetric flask, diluted with distilled water, and analyzed for PTE content.

Ethanol extracts (herbal tinctures) were prepared by mixing 5 g of ground dry plant material (shoots) with 30 mL of 70% ethanol at room temperature for 72 h. Samples were filtered afterward, and re-extracted with fresh solvent (20 mL). After 48 h of additional mixing, the suspension was filtered through filter paper while the supernatants were combined. Obtained solutions were evaporated to 5 mL, transferred into a 25 mL volumetric flask, diluted with distilled water, and analyzed for PTE content.

2.3. Determination of PTE Contents in Samples

PTE contents in all samples (soil, plants and extracts) were determined by using inductively coupled plasma optical emission spectroscopy (ICP-OES)—model iCAP 6500 (Thermo Fischer Scientific, Waltham, MA, USA). The analysis involved three replicates. Each of the replicates was analyzed by the instrument in triplicate. Operating parameters were tuned as follows: cooling flow speed—12 L min⁻¹, nebulizer gas flow speed—0.5 L min⁻¹, axial gas flow—0.5 L, and RF generator strength—1150 W.

Ten blank solutions prepared by the standard procedure were analyzed to determine the limit of quantification (LOQ) and limit of detection (LOD). The LOQ and LOD were calculated as three and ten times the standard deviation of blank solutions, respectively.

2.4. Validation of the Method (QA/QC)

The accuracy of the analytical procedure (the microwave digestion, acid mixture, temperature conditions) and precision of the instrument conditions were verified using certified reference materials: ERM-CD281 (ryegrass), provided by the Institute for Reference Materials and Measurements (Geel, Belgium), and SRM 2711a (Montana II soil), issued by the National Institute of Standards and Technology (NIST, Gaithersburg, MD, USA). Recovery values were between 90.3% and 102.7%.

2.5. Health Risk Assessment

2.5.1. Non-Carcinogenic Risk

The non-carcinogenic risk of PTEs in (sub)urban soils for children and adults was assessed by hazard index (HI) and hazard quotients (HQ) for different exposure pathways—ingestion (HQ_ing), dermal contact (HQ_der), and inhalation (HQ_inh). Cumulative values of HI and HQ for all PTEs were determined to assess the summary effect of all examined PTEs (CHI, CHQ_ing, CHQ_der, and CHQ_inh). Calculations were carried out according to the USEPA recommendations [27,28], using Equations (1)–(8), where C represents the determined pseudo-total concentrations of a PTE in the soil (mg kg⁻¹), while other parameters (exposure frequency—EF, exposure duration—ED, body weight—BW, inhalation unit risk—IUR, etc.) are given in Supplementary Table S3. HQ_inh was not determined for Fe and Sr, since their reference concentrations for inhalation are not defined. HQ and HI values below 1 indicate no adverse health effects [29]. Increased levels suggest an increased possibility of negative effects on human health.

$$\mathrm{HI}={\mathrm{HQ}}_{\mathrm{ing}}+{\mathrm{HQ}}_{\mathrm{der}}+{\mathrm{HQ}}_{\mathrm{inh}}$$

(1)

$$\mathrm{CHI}=\sum \mathrm{HI}$$

(2)

$${\mathrm{HQ}}_{\mathrm{ing}}={\displaystyle \frac{\mathrm{C} \times \mathrm{IngR} \times \mathrm{RBA} \times \mathrm{EF} \times \mathrm{ED}}{\mathrm{BW} \times \mathrm{AT} \times \mathrm{RfD}}\times {10}^{-6}}$$

(3)

$${\mathrm{CHQ}}_{\mathrm{ing}}=\sum {\mathrm{HQ}}_{\mathrm{ing}}$$

(4)

$${\mathrm{HQ}}_{\mathrm{der}}={\displaystyle \frac{\mathrm{C} \times \mathrm{SA} \times \mathrm{AF} \times \mathrm{ABS} \times \mathrm{EF} \times \mathrm{ED}}{\mathrm{BW} \times \mathrm{AT} \times \mathrm{RfD} \times \mathrm{GIABS}}\times {10}^{-6}}$$

(5)

$${\mathrm{CHQ}}_{\mathrm{der}}=\sum {\mathrm{HQ}}_{\mathrm{der}}$$

(6)

$${\mathrm{HQ}}_{\mathrm{inh}}={\displaystyle \frac{\mathrm{C} \times \mathrm{InhR} \times \mathrm{EF} \times \mathrm{ED}}{\mathrm{BW} \times \mathrm{AT} \times \mathrm{RfC} \times \mathrm{PEF}}}$$

(7)

$${\mathrm{CHQ}}_{\mathrm{inh}}=\sum {\mathrm{HQ}}_{\mathrm{inh}}$$

(8)

The estimated daily intake (EDI, mg kg⁻¹ day⁻¹) for children and adults and the non-carcinogenic risk (THQ) of PTEs in C. bursa pastoris and herbal extracts were assessed according to Equations (9) and (10), as per USEPA [29]. Total non-carcinogenic risk (TTHQ) was calculated as the sum of THQ for each PTE—Equation (11).

$$\mathrm{EDI}={\displaystyle \frac{\mathrm{IRD}\times \mathrm{C}}{\mathrm{BW}}}$$

(9)

$$\mathrm{THQ}={\displaystyle \frac{\mathrm{C} \times \mathrm{IRD} \times \mathrm{EF} \times \mathrm{ED}}{\mathrm{BW} \times \mathrm{AT} \times \mathrm{RfD}}}$$

(10)

$$\mathrm{TTHQ}=\sum \mathrm{THQ}$$

(11)

C in the equation refers to the concentration of a PTE in the raw herb and water and ethanol extracts prepared from the plant shoots (mg kg⁻¹). Other parameters (ingestion rate—IRD, averaging time—AT, reference dose—RfD, etc.) are given in Supplementary Table S3.

2.5.2. Carcinogenic Risk

The total carcinogenic risk (TCR) of PTEs in soils was calculated as a sum of carcinogenic risks (CR) through ingestion, dermal exposure and inhalation pathways for each examined PTE. CCR and CTCR represent the sum of CR and TCR for all PTEs, i.e., the cumulative influence of all PTEs present in the matrix. CR, TCR and their cumulative values were determined according to Equations (12)–(19).

$$\mathrm{TCR}={\mathrm{CR}}_{\mathrm{ing}}+{\mathrm{CR}}_{\mathrm{der}}+{\mathrm{CR}}_{\mathrm{inh}}$$

(12)

$$\mathrm{CTCR}=\sum \mathrm{TCR}$$

(13)

$${\mathrm{CR}}_{\mathrm{ing}}={\displaystyle \frac{\mathrm{C} \times \mathrm{IFS} \times \mathrm{RBA} \times \mathrm{CSFo}}{\mathrm{AT}}\times {10}^{-6}}$$

(14)

$${\mathrm{CR}}_{\mathrm{ing}}=\sum {\mathrm{CR}}_{\mathrm{ing}}$$

(15)

$${\mathrm{CR}}_{\mathrm{der}}={\displaystyle \frac{\mathrm{C} \times \mathrm{DFS} \times \mathrm{ABS} \times \mathrm{CSFo}}{\mathrm{AT} \times \mathrm{GIABC}}\times {10}^{-6}}$$

(16)

$${\mathrm{CCR}}_{\mathrm{der}}=\sum {\mathrm{CR}}_{\mathrm{der}}$$

(17)

$${\mathrm{CR}}_{\mathrm{inh}}={\displaystyle \frac{\mathrm{C} \times \mathrm{EF} \times \mathrm{ED} \times \mathrm{IUR} \times 1000}{\mathrm{AT} \times \mathrm{PEF}}\times {10}^{-6}}$$

(18)

$${\mathrm{CCR}}_{\mathrm{inh}}=\sum {\mathrm{CR}}_{\mathrm{inh}}$$

(19)

In the same way, lifetime carcinogenic risk (LCR) was evaluated for plant samples and prepared extracts—Equation (20). If CR, TCR, and LCR are below 1, no adverse effect on human health could be expected.

$$\mathrm{LCR}={\displaystyle \frac{\mathrm{C} \times \mathrm{IRD} \times \mathrm{EF} \times \mathrm{ED} \times \mathrm{CSFo}}{\mathrm{BW} \times \mathrm{AT}}\times {10}^{-6}}$$

(20)

Carcinogenic risks were quantified for those elements with defined slope factors CSFo (Cr and Ni) and are considered acceptable if calculated values are in the range 1 × 10⁻⁶ to 1 × 10⁻⁴ [30]. All parameters are given in Supplementary Table S3.

2.6. Geo-Accumulation Index, Bioconcentration and Translocation Factors

The geo-accumulation index was defined as follows [31,32]:

$${\mathrm{I}}_{\mathrm{geo}}={\log }_2\left( {\displaystyle \frac{C_{\mathrm{n}}}{1.5 \times {\mathrm{BC}}_{\mathrm{n}}}} \right)$$

(21)

where Cn represents the measured concentration of PTE “n” in the soil, and BC_n is the corresponding background concentration for the measured PTE in world soils. BC_n values were defined as the elemental background concentration in world soils [33,34].

To evaluate the C. bursa pastoris potential for PTE accumulation and translocation, bioconcentration and translocation factors were determined. The bioconcentration factor (BCF) was calculated as a ratio among PTEs contents in roots and the corresponding pseudo-total PTE concentrations in soil. The translocation factor (TF) represents the ratio between the PTE concentration in shoots and the concentration in roots.

2.7. Data Analysis

Descriptive statistics were used to summarize the distribution of PTE concentrations in soils, plants, and extracts. The Kolmogorov–Smirnov test assessed data normality to select the appropriate further statistical tests. Since the data did not fit to normal distribution, the non-parametric Kruskal–Wallis test (KW), followed by the post hoc Dunn test, was applied to identify significant differences in PTE levels across sample types. Principal component analysis (PCA) and Spearman correlation analysis were conducted to explore patterns and relationships among PTEs. All analyses were performed using R Studio, version 4.2.3.

3. Results and Discussion

3.1. PTE Contents in Soils

The pseudo-total concentrations of PTEs in soils (

Table 1. PTEs phytoavailable (EDTA) and pseudo-total (AR) content in soil samples (mg kg⁻¹).

			Ba	Co	Cr	Cu	Fe	Mn	Ni	Sr	Zn
EDTA	BG	mean	7.16 ab	1.41 ab	0.06 a	7.63 b	168.79 a	156.92 a	3.13 ab	16.46 a	15.82 b
		min	3.96	0.81	0.01	5.41	88.07	99.1	2.58	13.75	13.02
		max	9.26	2.17	0.13	10.15	269.6	243.76	3.46	20.69	17.27
		stdev	2.42	0.59	0.05	1.98	78.91	64.96	0.39	2.94	2.02
	SM	mean	2.85 b	0.55 ab	0.04 a	8.66 b	154.65 a	112.41 a	3.11 a	14.25 ab	18.50 b
		min	1.22	0.39	0	2.22	112.53	87.95	1.35	10.64	11.77
		max	4.3	0.66	0.12	12.37	220.55	142.03	4.51	20.41	25.29
		stdev	1.33	0.11	0.05	4.73	48.99	22.83	1.33	4.26	5.66
	VR	mean	11.19 a	1.80 a	0.10 a	11.97 ab	234.10 a	161.39 a	2.73 ab	12.02 ab	23.72 ab
		min	9.64	1.23	0.05	7.86	177.34	110.89	2.03	8.87	18.79
		max	12.48	2.82	0.16	14.5	266.71	232.61	3.82	16.07	26.57
		stdev	1.21	0.76	0.04	3.1	41.23	53.07	0.82	3.12	3.68
	BO	mean	3.91 ab	1.10 ab	0.08 a	279.74 a	252.20 a	140.88 a	0.90 b	10.68 b	85.87 a
		min	2.07	0.18	0	159.34	93.49	49.42	0.24	7.16	40.4
		max	5.17	1.65	0.18	440.01	335.31	194.32	1.31	12.65	174.6
		stdev	1.4	0.69	0.07	124.4	118.49	68.59	0.5	2.6	65.89
		p value	***	**	ns	***	ns	ns	***	*	***
AR	BG	mean	35.79 b	7.45 a	30.45 ab	16.21 b	14,902.54 a	334.13 a	26.02 ab	41.30 ab	49.70 b
		min	25.6	6.47	21.08	12.11	12,237.57	295.74	15.94	22.36	40.39
		max	42.12	9.27	41.33	22.35	18,468.96	397.37	38.99	60.66	65.23
		stdev	7.37	1.18	8.03	4.48	2735.76	42.21	9.98	16.2	11.63
	SM	mean	34.85 ab	8.10 a	42.69 a	21.11 ab	16,613.58 a	384.13 a	41.86 a	101.84 ab	66.50 ab
		min	10.72	6.77	37.91	8.42	14,773.04	358.53	31.92	69.16	63.2
		max	47.82	9.64	49.46	29.42	17,681.63	397.36	56.01	164.67	69.38
		stdev	17.77	1.15	4.49	9.46	1241.56	15.18	10.7	44.78	2.48
	VR	mean	62.43 a	7.05 a	25.69 b	21.55 ab	14,568.10 a	391.37 a	16.06 ab	25.18 b	56.23 ab
		min	52.43	4.39	20.03	15.11	9677.42	200.75	9.49	20.94	50.31
		max	67.93	8.66	30	26.35	18,839.49	574.74	23.29	29.91	67.94
		stdev	5.89	1.87	4.06	4.9	3876.18	156.76	5.86	3.61	8.58
	BO	mean	44.40 ab	8.18 a	26.77 ab	533.16 a	14,189.69 a	428.17 a	9.92 b	153.48 a	193.47 a
		min	27.7	3.82	21.26	341.51	11,391.85	175.56	7.3	111.8	90.52
		max	65.93	11.54	36.17	689.95	16,409.97	566.82	12.65	187.39	396.04
		stdev	16.23	3.25	6.57	151.53	2083.09	188.18	2.13	32.26	151.1
		p value	***	ns	***	***	ns	ns	***	***	***
European average			85.2	8.91	32.6	16.4	/	524	30.7	130	60.9
World average			19–2368	25	54	13–24	5%	437	22	87–210	64
Remediation values			625	240	30	75	/	/	210	/	720

European average—[23], world average—[24], remediation values—[25], min—minimum, max—maximum, stdev—standard deviation values; significance levels: *** p < 0.001, ** p < 0.01, * p < 0.05, ns—not significant. Different letters in each column indicate significant differences between sampling sites according to Dunn’s post hoc test. Values above the European average concentrations are in bold.

) exceeded the European (85.2, 8.91, 32.6, 16.4, 524,30.7, 130, 60.9 mg kg⁻¹ for Ba, Co, Cr, Cu, Mn, Ni, Sr and Zn, respectively) and world average PTE levels (Ba—19–2368 mg kg⁻¹, Co—25 mg kg⁻¹, Cr—54 mg kg⁻¹, Cu—13–24 mg kg⁻¹, Mn -437 mg kg⁻¹, Ni—22 mg kg⁻¹, Sr—87–210 mg kg⁻¹, Zn—64 mg kg⁻¹) [35,36] in at least one sampling point, most notably at BO1 and BO2. Particularly, the highest values (35.93 mg kg⁻¹ of Ba, 689.95 mg kg⁻¹ of Cu, 566.82 mg kg⁻¹ of Mn, 187.39 mg kg⁻¹ of Sr, and 396.04 mg kg⁻¹ of Zn) were observed in samples from city centers, while the peripheries held the lowest amounts (10.72, 3.82, 20.03, 8.42, 175.56, 111.8, and 22.36 mg kg⁻¹ for Ba, Co, Cr, Cu, Mn, Ni and Sr, respectively) (Table 1). The overall mean contents in soils followed the order Mn > Cu > Zn > Sr > Ba > Cr > Ni > Co (Table 1). Anthropogenic activities strongly affected PTEs fractions in soils as they varied significantly both across the sampling cities and within each site. The uneven distribution of PTEs and high variability may indicate the relationship between urbanization and its influence on soil composition [37].

Due to long-term atmospheric deposition from the former copper smelter, BO held extreme levels of Cu. Pseudo-total and EDTA-extractable concentrations exceeded twenty times the European average for Cu of 16.4 mg kg⁻¹ [23]. Moreover, they surpassed the remediation value of 75 mg kg⁻¹, defined by Serbian legislators [38], by seven times (Table 1). These results indicate that the main soil functions are hindered and soil remediation and recovery are necessary. Similarly, Zn concentrations exceeded three times the world and European mean (64 and 60.9 mg kg⁻¹, respectively), with the highest level of 193.47 mg kg⁻¹ also recorded in BO [35,36]. In particular, BO1, sampled close to the mine processing plant, contained 689 mg kg⁻¹ of Cu and 396.04 mg kg⁻¹ of Zn (Table 1). Both levels could be considered phytotoxic since they surpassed ten and two times the toxicity limits for Cu and Zn, defined as 60 mg kg⁻¹ and 200–300 mg kg⁻¹, respectively [39]. Additionally, BO1 and BO2 exceeded the European mean content for Sr (130 mg kg⁻¹) and Mn (524 mg kg⁻¹) [35], and were the most abundant with Co compared to other sampling sites (3.25 mg kg⁻¹) (Table 1). On the other hand, SM held the highest contents of Cr (42.69 mg kg⁻¹) and Ni (41.86 mg kg⁻¹). The maximum level of Ni was recorded in the alluvial plain soil (SM3) (Table 1). Ni is believed to be highly associated with ultramafic parent rocks [36,40], where the highest values were identified in soils developed from alluvium [36].

In general, Cu was the element with the highest concentration in BO, while BG, SM and VR held the highest abundance of Mn. Albeit generally considered a lithogenic element, as the second most abundant in the Earth’s crust, after Fe [36], Mn was found to be relatively mobile. Cu and Zn were highly phytoavailabile, with up to 60% and 40% of the pseudo-total contents, respectively. Moreover, up to 65% of the Sr pseudo-total was determined in the phytoavailable pool, as it can be easily mobilized during soil weathering [36]. A high availability of Sr directly leads to increased uptake by plants, potentially causing phytotoxicity and health risks for human health [41].

The I_geo index mirrored concentration levels in soils. BO had I_geo for Cu in a range of 4.25 to 5.26 (Figure 2d), demonstrating heavy to extreme contamination, the fifth and sixth categories as per Müller [31]. It also exhibited moderate to heavy contamination with Zn, with I_geo values ranging from 1.06 to 2.46 (Figure 2). An earlier study on soil contamination in BO also reported high I_geo values for these PTEs [42]. BG and SM were highly contaminated with Sr. I_geo indices signified moderate to heavy contamination, i.e., grades 2 and 3 (Figure 2a,b).

PTEs levels in soils followed the hypothesized pollution gradient. I_geo increased with decreasing vicinity to the city center, demonstrating moderate to heavy soil contamination with Ni, Sr, and Zn, and even extreme pollution with Cu in BO (Figure 2d). Similar findings have been reported in other urban environments, where elevated levels of Cu and Zn were linked to anthropogenic inputs such as industrial activity, traffic, and waste incineration [42,43,44].

Considering urban soils, geochemically originated Co in urban areas was found in Beijing, China [44], while low I_geo for Co was also reported in the soils of Havana, Cuba [43]. In another study on mining-impacted urban settings in Bor (Serbia), Cr also had I_geo < 0 and showed weak associations with the Cu–Zn group [42]. Conversely, I_geo > 2 for Cu was documented, accompanied by strong Cu–Zn–As correlations, while Ba and Fe maintained I_geo values around or below zero. In Baotou City, a major industrial city (Inner Mongolia, China), Cu and Zn also emerged as primary contaminants, with I_geo values ranging from low to heavily polluted categories [37]. Our results are consistent with earlier studies conducted in industrial centers in Serbia, including Pančevo, Smederevo, and Obrenovac, which reported excessive Cu and Zn as a result of anthropogenic activities [45].

3.2. PTE Contents in C. bursa-pastoris

PTE concentrations in C. bursa-pastoris underground and aerial parts (

Table 2. PTE contents in roots and shoots of C. bursa-pastoris (mg kg⁻¹).

			Ba	Co	Cr	Cu	Fe	Mn	Ni	Sr	Zn
Roots	BG	mean	26.33 a	0.43 a	12.85 a	5.10 ab	792.52 a	35.91 a	7.90 a	40.03 ab	39.24 ab
		min	14.6	0.05	2.37	3.23	201.85	14.88	3.68	17.97	30.49
		max	35.19	0.79	18.59	6.51	1387.47	54.34	10.89	52.9	44.21
		stdev	9.05	0.32	7.84	1.43	510.56	17.02	3.22	16.54	6.49
	SM	mean	17.48 a	0.30 a	9.04 a	5.99 ab	566.93 a	29.74 a	5.90 a	34.21 ab	37.46 ab
		min	11.58	0.12	6.07	4.22	271.39	18.76	4.61	30.96	32.82
		max	21.02	0.42	14.33	7.94	799.19	36.72	7.46	37.89	45.96
		stdev	4.39	0.13	3.87	1.57	229.49	8.2	1.17	2.82	6.34
	VR	mean	20.89 a	0.16 ab	5.58 ab	4.57 b	547.11 a	22.47 a	3.25 ab	27.18 b	30.39 b
		min	12.42	0.1	3.72	4.14	330.89	17.18	2.16	23	23.54
		max	25.54	0.22	7.53	5.35	728.52	26.23	4.28	31.37	39.92
		stdev	6.27	0.04	1.5	0.52	171.65	4.01	0.89	3.51	7.17
	BO	mean	17.64 a	0.09 b	2.54 b	23.36 a	353.93 a	24.32 a	1.68 b	57.04 a	68.98 a
		min	15.84	0.02	1.53	19	158.93	13.85	0.45	36.18	37.12
		max	19.75	0.14	4.1	30.74	512.1	32.72	3.22	69.22	86.37
		stdev	1.6	0.05	1.12	5.53	152.44	8.16	1.2	15.2	23.84
		p value	ns	*	**	***	ns	ns	***	***	**
Shoots	BG	mean	20.62 ab	0.27 a	9.27 a	5.57 ab	480.00 a	30.75 a	6.26 a	49.56 ab	45.48 a
		min	19.57	0.19	5.07	4.79	330.31	27.18	4.64	30.85	30.41
		max	21.81	0.38	13.9	7.02	604.98	33.01	9.17	59.58	55.98
		stdev	0.88	0.07	3.76	1.08	119.46	2.58	2.16	13.69	11.37
	SM	mean	14.97 ab	0.15 ab	5.19 a	5.47 ab	343.61 ab	29.24 ab	6.03 a	42.67 ab	30.74 a
		min	13.98	0.03	4.63	3.79	309.7	21.53	2.06	29.06	22.94
		max	16.89	0.23	5.98	6.58	400.51	34.2	13.15	52.72	36.69
		stdev	1.36	0.09	0.58	1.25	40.66	5.74	5.29	10.23	6
	VR	mean	19.60 a	0.14 ab	3.54 ab	5.40 b	364.11 ab	26.12 ab	1.87 ab	39.21 b	33.84 a
		min	12.5	0.09	2.94	3.71	302.29	22.35	1.15	34.28	30.98
		max	23.58	0.18	4.33	6.97	486.55	31.3	2.76	45.64	35.87
		stdev	5.21	0.03	0.56	1.37	89.61	3.88	0.69	4.91	2.11
	BO	mean	13.11 b	0.05 b	0.27 b	20.21 a	175.28 b	21.61 b	0.63 b	60.78 a	49.99 a
		min	6.97	0.02	0.09	10.48	123.66	15.71	0.34	39.51	32.81
		max	22.02	0.1	0.58	34.05	215.31	24.91	0.94	81.59	67.81
		stdev	6.71	0.04	0.22	10.56	39.46	4.31	0.23	17.78	15.04
		p value	*	***	***	***	***	**	***	*	ns

min—minimum, max—maximum, stdev—standard deviation values; significance levels: *** p < 0.001, ** p < 0.01, * p < 0.05, ns—not significant. Different letters in each column indicate significant differences between sampling sites, according to Dunn’s post hoc test.

) suggested the plant primarily absorbed the labile and highly available PTEs in soil.

As shown in Table 2, the Zn content in underground tissues ranged from 23.06 to 67.62 mg kg⁻¹. Although predominantly accumulated in roots, almost 91% of Zn_roots on average was transferred to the shoots. The concentration range in shoots was 23.29–86.28 mg kg⁻¹, with the highest levels recorded in samples from the most urbanized sites, BO and BG. Conversely, samples from suburban zones contained the lowest Zn abundance. These results are in accordance with previous reports. An average content of 88.8 mg kg⁻¹ of Zn was found in the roots of C. bursa-pastoris from the urban park of the Botanical Garden of Komarov Botanical Institute [46], while C. bursa-pastoris leaves sampled in rural and urban areas of Bradford, UK, contained between 3.5 and 33.2 mg kg⁻¹, where the highest values were determined in urban areas [47].

Likewise, Sr concentration in roots ranged from 18.06 mg kg⁻¹, determined in the BG3, to 68.81 mg kg⁻¹ found in BO1. In general, Sr is not easily transported from underground to aerial plant parts [36]. Reported concentrations in shoots were considerably lower, and had a range of 29.31–81.06 mg kg⁻¹, as seen in Table 2, reaching its maximum in BO. As previously reported, plants could exhibit toxicity signs at levels above 30 mg kg⁻¹ [41]. Besides Zn and Sr, C. bursa-pastoris actively accumulated Mn in roots. The concentrations in roots ranged from 22.47 mg kg⁻¹ found in VR to 35.91 mg kg⁻¹ determined in BG (Table 2). Mn content in shoots varied significantly, with the lowest value determined in samples from BO (15.88 mg kg⁻¹) and the highest in SM (34.04 mg kg⁻¹). Despite relatively low concentrations in soil, C. bursa-pastoris significantly absorbed Cr in roots at all sites. The Cr content in C. bursa-pastoris surpassed the WHO permissible limit in plants of 1.50 mg kg⁻¹ [48]. In general, samples from Belgrade contained the highest levels of Cr in both roots and shoots (18.57 mg kg⁻¹ and 13.77 mg kg⁻¹, respectively).

Contrary to Sr and Zn, Cu was principally translocated to the aboveground parts, which are commonly used for extract preparation and could potentially cause the risk of consumption or dermal application. According to WHO, the permissible limit of Cu in edible plants is 3.00 mg kg⁻¹ [49]. Concentrations in C. bursa-pastoris exceeded this limit (Table 2). However, the WHO limits for medicinal herbs have not been established yet [50]. The maximum levels of Cu in roots and shoots were 30.69 and 33.90 mg kg⁻¹, respectively, determined in BO1 (Table 2). Ni and Co also exhibited high mobility within the plant. The highest average content of Ni was identified in BG shoots, with significant differences among sites. These values were up to five times higher than the Ni permissible limit in herbs of 1.63 mg kg⁻¹ [1].

Differences in PTE accumulation and mobility patterns were reflected in bioconcentration (BCF) and translocation factors (TF). Zn, which was highly abundant in both roots and shoots, had a BCF value of 5.31 (BO2), and a maximum calculated TF of 2.24 (BO3) (Supplementary Tables S4–S6). Sr was mainly accumulated in roots with BCF varying from 0.91 to 5.41. Additionally, coefficients for Mn ranged from 1.47 in BO to 6.31 in BG. On the other hand, Cu, Ni, and Co, mainly transferred to the shoots, showed high TF levels. TF_Cu ranged from 0.92 to 1.22, TF_Ni from 1.03 to 1.23, and TF_Co had average values above 1 at all sampling sites.

Overall, plant material sampled in more industrially developed cities was more contaminated with PTEs. BO contained the highest levels of Cu, Sr, and Zn, while BG held the greatest amounts of several PTEs—Ba, Co, Cr, Fe, Mn, and Ni. The obtained results suggested that C. bursa-pastoris can store PTEs in roots and shoots, demonstrating two strategies in PTEs uptake—exclusion, i.e., element accumulation in roots, and accumulation, i.e., transfer to the aerial parts. Ba, Cr, Mn, and Sr were mainly deposited in the roots, while Cu, Ni, Co, and Fe were predominantly translocated to the shoots. Zn exclusively demonstrated both strategies.

Elevated PTE contents in plants could have detrimental effects on plant growth [7]. Even beneficial elements in overdose could impair the processes required for basic plant functions and survival. For instance, an excessive amount of Cu may cause adverse effects on plant growth, photosynthesis, germination, and the antioxidant response [39]. Likewise, plants growing in Zn-contaminated soils may exhibit signs of toxicity, such as growth inhibition, mineral nutrition deviations, reduced respiratory and photosynthetic rates, and an elevated generation of reactive oxygen species (ROS) [7]. Sr chemical toxicity may influence adverse effects on plant growth, chromosomal abnormalities, induced oxidative stress, and nutritional disbalance, as Sr may replace Ca owing to their chemical similarities [41]. Mn toxicity symptoms include oxidative stress, disrupted photosynthesis, inhibited chlorophyll biosynthesis, and prevented nutrient uptake and translocation [51,52]. Cr may influence increased ROS generation, disrupt essential plant metabolic processes, and impair cellular integrity while attacking DNA, membrane proteins, and lipids [53].

Roots and shoots are consumable parts of C. bursa-pastoris. Thus, the elemental content of prepared herbal infusions and tinctures could be susceptible to contamination.

3.3. Correlation Analysis

Correlation analysis revealed significant interconnection among PTEs’ mobility in the soil–plant system and their migration within the plant (Figure 3a). Correlations between PTE contents in the plant and soil fractions suggested the crucial influence of the content and availability in soil on the accumulation in C. bursa-pastoris. Despite the significant differences found among examined cities, Zn in plants strongly correlated with the pseudo-total Zn fraction in soil. These findings align with the results reported by Drozdova et al. [46], demonstrating that the accumulation potential of C. bursa-pastoris reflects the concentration levels of PTEs recorded in the soil. Furthermore, both Cu_roots and Cu_shoots showed a significant correlation with Cu_EDTA and pseudo-total phases. Likewise, Ni_roots highly correlated with Ni_EDTA and Ni_pseudo-total. Furthermore, strong correlations between under- and aboveground parts confirmed the translocation ability of C. bursa-pastoris. In particular, Zn_roots strongly correlated with Zn_shoots (0.52), as did Cu_roots with Cu_shoots (0.66) (Figure 3a). A significant correlation was also identified among Ni_roots and Ni_shoots (0.72), proving the presence of element transfer through plant tissues. Sr in plants significantly correlated with Zn and Cu, reflecting the relationship between these elements in soils and their joint absorption.

Correlations among pseudo-total and EDTA-extractable fractions of Cu (0.96), Mn (0.51) and Zn (0.58), indicated a potential mutual contamination origin (Figure 3a). Together with this evidence, high mobility and strong correlations could indicate anthropogenic sources in urban environments, such as municipal waste, abrasion dust, construction materials, and industrial emissions [43]. Sr_AR highly correlated with Cu_AR and Zn_AR, which could suggest a shared anthropogenic source, potentially linked to urban pollutants or localized industrial activities. The strong linkage of Co, Cr, Ba, and Ni with Fe (0.58 < ρ < 0.83), low availability, and low I_geo index suggested their lithogenic origin.

3.4. PCA Analysis

PCA analysis allowed for the identification of site-specific elemental signatures and highlighted key contributors to spatial variability. It confirmed the distinction of samples from different sites based on the levels of PTEs (Figure 3b). Cu and Zn, both mobile and pseudo-total phase, as well as the Sr pseudo-total fraction, influenced the separation of the cluster belonging to BO along the PC1 component, suggesting that the BO site may be subject to anthropogenic inputs. The co-occurrence of mobile and pseudo-total forms of Cu and Zn further indicates a potential environmental risk, as these elements can be both abundant and bioavailable. In the same way, positive pseudo-total loadings of Cr and Ni, as the most abundant, caused the differentiation of scores belonging to SM. Thus, Cr and Ni could thus be considered principal drivers of the variability in the SM site. EDTA-extractable Ba affected the differentiation of the VR scores, since the mobile Ba fraction was dominant. The BG site slightly overlapped with SM and VR clusters, indicating a degree of similarity in elemental composition. This overlap may result from comparable PTE concentrations or shared sources of contamination, such as diffuse atmospheric deposition or similar geological substrates.

3.5. PTE Contents in Herbal Extracts

PTE contents in aqueous extracts followed a similar trend to concentrations in shoots. BO extracts showed the highest levels of Zn, Sr, Cu, and Ba (Supplementary Tables S7). Zn was the most abundant among PTEs in water and ethanol extracts, with concentrations ranging from 5.32 to 10.07 mg kg⁻¹. Up to 22% of Zn content in shoots was extracted during infusion processes, i.e., it could be considered accessible for humans. Nevertheless, the bioaccessibility of Cr and Ni, group 1 carcinogens, was up to 42 and 34% on average, respectively (Figure 4a). Although nearly 90% of Cr was extracted from BO shoots (Figure 4a), the concentration of Cr in herbal extracts did not exceed the limit of 2 mg kg⁻¹ defined for Cr in herbal medicines [49]. Ni concentrations in prepared extracts were acceptable, despite exceeding permissible limit in raw herbs.

Ethanol was found to be a much weaker solvent than water for PTE extraction, since only five elements (Zn, Sr, Mn, Cu, and Ba) were detected in the alcohol extracts, with others being below the detection limit (Supplementary Tables S7 and S8). The highest levels of bioaccessibility (%) in ethanol extracts were recorded for Cu and Zn (Figure 4b). These two elements were also the most bioaccessbile, although significantly less than in water extracts.

Herbal product contamination could lead to PTEs entering into the food chain and posing severe implications for human well-being.

3.6. Health Risk Assessment

3.6.1. Risk Assessment of PTEs in Soils

Soil hazard quotients (HQ) for individual elements were calculated for children and adults for ingestion (HQ_ing), dermal contact (HQ_der), and inhalation exposure pathways (HQ_inh). HQ and hazard index (HI), and their cumulative levels (CHQ and CHI), were within the acceptable limit (<1) (Supplementary Table S9). As per exposure assessment, values obtained for PTEs decreased in the following order, HQ_ing > HQ_der > HQ_inh, suggesting that the risk of an accidental increase in PTE soil content could be posed in the ingestion route. The greatest HQ_ing for both children and adults was noticed for Fe, Mn, and Cr in all four cities. Fe and Mn were the most abundant elements in soils, probably originating from the parent material. Our results indicate that the highest risk for children could potentially occur through the ingestion of Cr-contaminated soils, while Ba was found to be the most hazardous out of nine PTEs in cases of inhalation by children. As for adults, HQ_inh was the greatest for Mn. In general, the highest CHQ_ing and CHQ_inh were determined in BO. The greatest CHQ for children and adults’ dermal exposure was calculated for SM.

Carcinogenic risk (CR) for all three exposure pathways was in the acceptable range (1 × 10⁻⁴ to 1 × 10⁻⁶) and could thus be considered risk-free (Supplementary Table S10). Cumulative CR (CCR) values for different exposure pathways decreased in the following order—CCR_ing > CCR_der > CCR_inh. Interestingly, the highest total CR (TCR) was calculated for soils from Sremska Mitrovica, the least populated and urbanized, and thus initially deemed the least polluted.

3.6.2. Risk Assessment of PTEs in Plants

The non-carcinogenic risk associated with the consumption of C. bursa pastoris was evaluated based on estimated daily intake (EDI) and target hazard quotient (THQ) for children and adults (Supplementary Tables S11 and S12). Owing to the higher element content in the raw herb compared to the extracts, the values of EDI (the highest values were 2.1 × 10⁻¹ for Fe_roots and 1.3 × 10⁻¹ for Fe_shoots) and THQs (8 × 10⁻⁵ and 1.4 × 10⁻⁵ for roots and shoots, respectively) were higher in C. bursa-pastoris than in its infusions and tinctures. Previous reports on risk assessment sensitivity and the effects of different parameters highlighted that the risk is most affected by PTE concentration [54]. In general, children were more vulnerable to health risks compared to adults. The highest EDIs of PTEs for both children and adults in raw roots and shoots were determined for Fe (roots: 2.1 × 10⁻³ for children and 5 × 10⁻² for adults, shoots: 1.3 × 10⁻¹ and 3 × 10⁻² for children and adults, respectively), followed by Sr (roots: 1.1 × 10⁻² for children and 2.5 × 10⁻³ for adults, shoots: 1.3 × 10⁻² and 3.1 × 10⁻³ for children and adults, respectively), and Zn (roots: 1 × 10⁻² for children and 2.5 × 10⁻³ for adults, shoots: 1.2 × 10⁻² and 2.8 × 10⁻³ for children and adults, respectively). The low bioaccessibility of Fe, Sr, and Zn might influence such a difference. Average EDIs and THQs for different cities were below the provisional tolerable daily intake (PTDI) established by the joint FAO/WHO expert committee on food additives [49].

3.6.3. Risk Assessment of PTEs in Herbal Extracts

EDI_children and EDI_adults for herbal infusion had values in the following order—Zn > Sr > Ba > Cu > Mn > Ni > Fe > Cr > Co (Figure 5a,b). Values for ethanol extracts followed the order Zn > Sr > Cu > Mn > Ba for children and Zn > Cu > Sr > Ba > Mn for adults (Supplementary Table S14). Values were below the established PTDIs, whereby the risk for children’s health was higher, as in the case of soil and raw herbs. EDI and THQ also indicated a higher risk from the consumption of water infusions, i.e., herbal teas, than tinctures (ethanol extracts) (Supplementary Tables S14–16). TTHQ for children and adults showed no significant difference regarding the sampling site. Although acceptable, the primary non-carcinogenic risk for infusions was calculated for Cr, with the maximum level VR. The highest THQ for tinctures was found for Cu. Due to the heavy Cu pollution, samples from Bor exhibited the greatest risk.

LCR values for teas and tinctures were in the acceptable range. The results suggested Ni as the highest contributor to the carcinogenic risk, with the highest levels found in BG, followed by BO, SM, and VR, respectively (Supplementary Figure S1).

Lifetime carcinogenic risks (LCR) for children and adults were below the acceptable risk range, i.e., 1 × 10⁻⁶ to 1 × 10⁻⁴, considering raw herbs (Supplementary Table S13).

4. Conclusions

The present study evaluated the concentrations and phytoavailability of PTEs in urban soils from four cities in Serbia, as well as their accumulation in the medicinal plant C. bursa-pastoris, comprehensively assessing the associated health risks. The results suggested that the contamination levels of soils from these urban areas were mainly influenced by anthropogenic activities. Soil analysis revealed two distinguished PTEs groups—highly abundant, mobile, and mutually correlated Cu, Zn, Sr, and Mn, and less available Co, Cr, Ba, and Ni. Despite heavy soil contamination with Ni, Sr, Cu, and Zn, the overall health risk assessment for different exposure pathways fell within acceptable ranges. While Sr, Cu, and Ni contents in the plant exceeded the safe limits, consumption of C. bursa-pastoris as a raw plant, herbal tea, or tincture was found to be safe, according to the target hazard quotient (THQ), estimated daily intake (EDI), and lifetime carcinogenic risk (LCR). The results demonstrated the higher sensitivity of children to hazardous effects in comparison to adults. These findings emphasize the need for integrated environmental and public health strategies to ensure the safe use of medicinal herbs. The present study highlights the importance of continuously monitoring PTE concentrations and mobility in soil–plant systems, as well as regulatory policy implementation in facing the challenges of PTE contamination of medicinal herbs. Future research should focus on the long-term monitoring of PTEs in medicinal plants across diverse environments and explore mitigation strategies for their contamination.