Comparison of the surgical resection and infarct 5/6 nephrectomy rat models of chronic kidney disease

Abstract

The 5/6 nephrectomy rat remnant kidney model is commonly used to study chronic kidney disease (CKD). This model requires the removal of one whole kidney and two-thirds of the other kidney. The two most common ways of producing the remnant kidney are surgical resection of poles, known as the polectomy model, or ligation of superior and inferior segmental renal arteries, resulting in pole infarction. These models have much in common, but also major phenotypic differences, and thus respectively model unique aspects of human CKD. The purpose of this review is to summarize phenotypic similarities and differences between these two models and their relation to human CKD while emphasizing their vascular phenotype. In this article, we review studies that have evaluated arterial blood pressure, the renin-angiotensin-aldosterone-system, autoregulation, nitric oxide, single-nephron physiology, angiogenic and antiangiogenic factors, and capillary rarefaction in these two models. In terms of phenotypic similarities, both models spontaneously develop hallmarks of human CKD including uremia, fibrosis, capillary rarefaction, and progressive renal function decline. They both undergo whole organ hypertrophy, hyperfiltration of functional nephrons, reduced renal expression of vascular endothelial growth factor, increased renal expression of antiangiogenic thrombospondin-1, impaired renal autoregulation, and abnormal vascular nitric oxide physiology. In terms of key phenotypic differences, the infarction model develops rapid-onset, moderate to severe systemic hypertension and the polectomy model develops early normotension followed by mild to moderate hypertension. Rats subjected to the infarction model have a markedly more active renin-angiotensin-aldosterone system. Comparison of these two models facilitates understanding of how they can be used for studying CKD pathophysiology.

INTRODUCTION

Chronic kidney disease (CKD) is the loss of kidney function, often irreversible, that occurs over a time of months or years. CKD may be characterized by reduced glomerular filtration rate (GFR), uremia, proteinuria, renal fibrosis, capillary rarefaction, metabolic acidosis, and skeletal myopathy and ultimately leads to organ failure (1–3). Approximately 15% of United States adults have CKD (4). The rat remnant kidney model of CKD is widely used experimentally due to its relative ease of generation and recapitulation of many human CKD characteristics. The most common rat remnant kidney model is the 5/6 nephrectomy (5/6Nx), in which one whole kidney is removed and the poles of the remaining kidney are ablated. Pole ablation is most commonly performed in one of two ways: surgically (e.g., removal by scalpel) (1, 5, 6) or infarction by ligation of the inferior and superior segmental arteries (Fig. 1, A and B) (6). The sham procedure includes laparotomy and renal pedicle manipulation (Fig. 1C). Although infarction and polectomy rats are both models of chronic renal insufficiency, they have important phenotypic differences and thereby model unique aspects of human CKD. Although the primary renal insult that initiates CKD in infarction and polectomy rats differ from typical human CKD, both rat models exhibit similar pathological progression from that point on. This progression includes compensatory adaptation of the remaining nephrons that, over time, leads to further damage, thus establishing a progressive cycle that may culminate with organ failure (Fig. 1D).

Figure 1.

Schematic summary of 5/6 nephrectomy (5/6Nx) by surgical resection (A), 5/6Nx by infarction (B), and sham-operated nephrectomy (C). All three procedures include kidney access by laparotomy. D: although the mechanism of primary renal insult is variable in human chronic kidney disease and different from both the infarction and polectomy 5/6Nx models, progressive decline of renal function is common to all. R, right.

The infarction and polectomy models recapitulate hallmarks of human CKD. Both models spontaneously develop elevated serum creatinine (5–12). Importantly, several studies have reported similar serum creatinine levels between infarction and polectomy rats 3 days following surgery (6, 13–15), suggesting a similar initial reduction of nephron number. Both models become uremic (16–20) and develop proteinuria (5, 6, 17, 21). Direct comparison studies have indicated that infarction rats have more severe proteinuria than polectomy rats through 6 wk postsurgery (Table 1) (6, 14, 15). For example, Ibrahim and Hostetter (15), at 2 wk postsurgery, found 154 mg/24 h of protein in polectomy rat urine compared with 249 mg/24 h in infarction rat urine. At 6 wk postsurgery, Griffin et al. (6) reported ∼10 mg/24 h of protein in polectomy rat urine compared with ∼50 mg/24 h in infarction rat urine. Studies that compared glomerular injury between infarction and polectomy rats have reported no difference (8, 15) or significantly more in infarction rats (6, 14). Both models develop some degree of tubulointerstitial injury and fibrosis (17, 22–26). Fibrosis, glomerular injury, and tubulointerstitial injury are virtually universal outcomes in human CKD (3, 27–29).

Table 1.

Proteinuria values for infarction and polectomy rats subjected to 5/6 nephrectomy from studies that compared these two models directly in a prospective manner

Week Postsurgery	Average Proteinuria, mg/24 h		References
	Infarction Model	Polectomy Model
2	249 ± 124 (±SD)	154 ± 73 (±SD)*	Ibrahim and Hostetter (15)
4	225 ± 93 (±SD)	135 ± 67 (±SD)*
3	11 ± 6 (±SE)	6 ± 0.6 (±SE)*	Griffin et al. (14)
4	∼35 ± 8 (±SE)	∼10 ± 3 (±SE)*	Griffin et al. (6)†
6	∼50 ± 10 (±SE)	∼10 ± 3 (±SE)*

These studies were performed in Sprague–Dawley rats. SD, standard deviation; SE, standard error; ∼, estimated values.

^*Significantly different than infarction rats from the corresponding time point; †proteinuria values for this table were estimated from a graphical figure in the referenced study.

Although these two ubiquitous rat models have much in common and both ultimately develop renal failure, they also have major phenotypic differences [e.g., degree of renin-angiotensin-aldosterone system (RAAS) activation, as detailed elsewhere] and therefore model unique aspects of human CKD. These differences may be underappreciated by the scientific community. Many published 5/6Nx rat studies 1) do not specify which 5/6Nx model they used, 2) lack rationale for model selection, and 3) omit discussion of their respective model’s unique pathophysiology. The aforementioned omissions may contribute to misinterpretation of results and inappropriate 5/6Nx rat publication referencing. Therefore, the goal of this review article is to raise awareness of the similarities and differences of these respective model phenotypes and aid investigators in model selection. Given that the kidney’s central function is blood filtration and the prominence of renal vasculature pathology in renal disease, we emphasize comparison of the infarction and polectomy rat renal vascular phenotypes. Because this topic is too large to comprehensively review, we compiled “mini-reviews,” each focused on a specific facet of vascular pathology. These include arterial blood pressure, the RAAS, autoregulation, nitric oxide (NO), single-nephron physiology, angiogenic and antiangiogenic factors, and capillary rarefaction.

In this review, we focus on assessing the variable of surgical approach (infarction vs. polectomy) upon multiple phenotypes. All studies referenced here used Sprague–Dawley or Wistar rat strains. Because the overwhelming majority of published studies used male rats, the majority of data we present are for male rats. However, recent studies have identified a number of phenotypic differences between males and females. We address these in sex differences in the 5/6nx model. In addition, the 5/6Nx surgery is sometimes performed as two separate surgeries. For example, a uniphrectomy performed at one surgery, and then sometime later, the poles of the remaining kidney are removed at a second surgery. The two-surgical step approach has been used for both polectomy (30–32) and infarction rats (12, 33). Although it is clear that both the one-step and two-step models manifest classical markers of CKD, including elevated phosphocreatine, fibrosis, proteinuria, etc., it is difficult to clearly determine differences between one-step and two-step surgeries for other phenotypes of interest. For example, in some reports, the first surgical step is pole ablation and the contralateral kidney is removed as a second surgery at a later time (32, 34, 35); other times, the reverse is used (36, 37). Within the body of two-surgical step 5/6Nx rat studies, there is variation in the amount of time elapsed between surgical steps (30, 32, 34–39), and there is variation in the surgical order. The effects of these variables upon the CKD phenotype are difficult to assess. Therefore, this review will focus on one-surgical step infarction and polectomy rats. Although we focus on reviewing the one-surgical step model, some two-surgical step publications are referenced. For these studies, we explicitly communicate, in text, that they are two-surgical step publications.

ARTERIAL BLOOD PRESSURE

Elevated blood pressure, as most recently defined by the American College of Cardiology and American Heart Association, is defined as a systolic blood pressure (SBP) of 120–129 mmHg and diastolic blood pressure (DBP) of <80 mmHg, stage 1 hypertension as a SBP of 130–139 mmHg or DBP of 80–89 mmHg, and stage 2 hypertension as a SBP of ≥140 mmHg or DBP of ≥90 mmHg (40). Hypertension affects ∼30% of United States adults (41). People with CKD are disproportionately hypertensive (42). Although there is no doubt that hypertension can contribute to CKD pathogenesis and progression, treatment of hypertension in people with CKD slows progression without abrogating it (43, 44), and not all people with CKD are hypertensive (42, 45). Thus, hypertensive and nonhypertensive CKD models are necessary to reflect different clinical manifestations of CKD.

Infarction and polectomy rats have markedly different arterial blood pressure phenotypes. The infarction model is characterized by rapid and robustly increased SBP (6, 8, 14, 15, 46, 47), whereas the polectomy model, in general, is characterized by early normotension, followed by slowly developing (over months) mild to moderate hypertension. For example, Griffin et al. (6), using telemetry-based measurements, found an infarction rat SBP of ∼160 mmHg with 3–5 days of surgery when both polectomy and sham-operated rats were normotensive. Many studies have reported later-onset (>6 wk postsurgery) hypertension in polectomy rats, typically 125–150 mmHg of SBP (5, 8, 17, 48, 49). In infarction rats, SBPs of >180 mmHg are common (20, 21, 25, 34, 50–54). Infarction and polectomy rat arterial blood pressures respond to angiotensin-converting enzyme (ACE) inhibitors and Ca²⁺ channel blockers (9, 35, 47–50, 52, 55–57). Significant changes in arterial blood pressure in response to experimental modulation of diet salt content have been reported in polectomy (58–60) and infarction rats (20, 21). In contrast, in infarction rats, Daniels and Hostetter (61) reported no statistically significant change in mean arterial blood pressure in high, low, and deficient (no salt) diet-fed relative to normal diet-fed controls. Table 2 shows SBP data for polectomy and infarction rats.

Table 2.

Summary of arterial blood pressure phenotype from infarction and polectomy rats subjected to 5/6 nephrectomy

Method of Measurement	Animal Sex	Animal Strain	Time Postsurgery	Average Systolic Blood Pressure, mmHg		References
				Infarction	Polectomy
Radiotelemetry of the abdominal aorta	Male	Sprague–Dawley	1 wk	∼160 ± 5 (±SE)	∼125 ± 3 (±SE)	Griffin et al. (6)*
			3 wk	∼155 ± 5 (±SE)	∼120 ± 3 (±SE)
			6 wk	∼155 ± 5 (±SE)	∼125 ± 3 (±SE)
Tail-cuff sphygmomanometer	Male	Sprague–Dawley	2 wk	169 ± 8 (±SE)	123 ± 6 (±SE)	Garber et al. (8)
			6 wk	169 ± 7 (±SE)	129 ± 9 (±SE)
Radiotelemetry of the abdominal aorta	Male	Sprague–Dawley	3 days	∼160 ± 10 (±SE)	∼130 ± 5 (±SE)	Griffin et al. (14)*
			1 wk	∼170 ± 10 (±SE)	∼120 ± 5 (±SE)
			3 wk	∼170 ± 10 (±SE)	∼120 ± 5 (±SE)
Tail-cuff sphygmomanometer	Male	Sprague–Dawley	2 wk	166 ± 31 (±SD)	120 ± 16 (±SD)	Ibrahim and Hostetter (15)
			4 wk	161 ± 28 (±SD)	120 ± 12 (±SD)
Carotid artery using a pressure transducer catheter	Male	Sprague–Dawley	2 wk		126 ± 9 (±SE)	Chuppa et al. (5)
			4 wk		121.7 ± 7 (±SE)
			5 wk		139.6 ± 4 (±SE)
			7 wk		135 ± 5 (±SE)
Tail-cuff sphygmomanometer	Male	Sprague–Dawley	Baseline		95.4 ± 10.8 (±SD)	Kang et al. (17)
			2 wk		100.8 ± 15.4 (±SD)
			4 wk		104.3 ± 15.4 (±SD)
			8 wk		145.3 ± 25.8 (±SD)
Radiotelemetry of the abdominal aorta	Male	Sprague–Dawley	14-15 wk		128.9 ± 2.3 (±SE)	Griffin et al. (48)
Radiotelemetry of the abdominal aorta	Male	Sprague–Dawley	1 wk	160 ± 5 (±SE)		Griffin et al. (9)*
			3 wk	160 ± 5 (±SE)
			6 wk	185 ± 5 (±SE)
Tail-cuff sphygmomanometer	Male	Sprague–Dawley	Baseline	∼110 ± 5 (±SE)		Sun et al. (25)*
			4 wk	∼180 ± 10 (±SE)
			8 wk	∼190 ± 10 (±SE)
Radiotelemetry of the abdominal aorta	Male	Sprague–Dawley	1 wk	∼165 ± 4 (±SE)		Griffin et al. (13)*
			2 wk	∼160 ± 4 (±SE)
			3 wk	∼160 ± 4 (±SE)
			4 wk	∼162 ± 4 (±SE)
			5 wk	∼170 ± 8 (±SE)
			6 wk	∼180 ± 8 (±SE)
			7 wk	∼182 ± 8 (±SE)
Carotid artery using a pressure transducer catheter	Male	Wistar	3 wk		115 ± 0.5 (±SE)	Juncos et al. (62)
			4 wk		116 ± 1.5 (±SE)
Tail-cuff sphygmomanometer	Male	Wistar	2 wk		139 ± 4 (±SE)	Sampaio-Maia et al. (63)
Tail-cuff sphygmomanometer	Male	Wistar	3 wk		∼122 ± 3 (±SE)	Bai et al. (64)*†
			5 wk		∼138 ± 4 (±SE)
			9 wk		∼140 ± 5 (±SE)
			11 wk		∼139 ± 4 (±SE)
Tail-cuff sphygmomanometer	Male	Wistar	9 wk (60 days)	191.3 ± 4.2 (±SD)		Correa et al. (16)
Tail-cuff sphygmomanometer	Male	Wistar	2 wk	∼190 ± 10 (±SE)		Lax et al. (21)*
			4 wk	∼195 ± 10 (±SE)
			6 wk	∼200 ± 10 (±SE)
			8 wk	∼220 ± 10 (±SE)
Tail-cuff sphygmomanometer	Male	Wistar	Baseline	124 ± 3 (±SE)		Vavrinec et al. (46)
			12 wk	176 ± 12 (±SE)
Tail-cuff sphygmomanometer	Male	Wistar	Baseline	∼125 ± 5 (±SE)		Vettoretti et al. (47)*
			6 wk	∼165 ± 5 (±SE)
			15 wk	∼170 ± 5 (±SE)
Tail-cuff sphygmomanometer	Male	Wistar	2 wk	∼175 ± 5 (±SE)		Anderson et al. (50)*
			4 wk	∼160 ± 10 (±SE)
			6 wk	∼170 ± 5 (±SE)
			8 wk	∼190 ± 5 (±SE)
			10 wk	∼185 ± 5 (±SE)
			12 wk	∼210 ± 5 (±SE)
Tail-cuff sphygmomanometer	Male	Wistar	3 wk	∼175 ± 5 (±SE)		Mackenzie et al. (52)*
			6 wk	∼185 ± 5 (±SE)
			9 wk	∼210 ± 5 (±SE)
			12 wk	220 ± 15 (±SE)
Tail-cuff sphygmomanometer	Male	Wistar	2 wk	∼170 ± 5 (±SE)		Anderson et al. (55)*
			3 wk	∼160 ± 10 (±SE)
			4 wk	∼170 ± 15 (±SE)
			5 wk	∼175 ± 10 (±SE)
			6 wk	∼175 ± 10 (±SE)
			7 wk	∼170 ± 15 (±SE)
			8 wk	∼180 ± 5 (±SE)
Tail-cuff sphygmomanometer	Male	Sprague–Dawley	3–4 wk	170.6 ± 7.9 (±SE)		Bidani et al. (65)
			6–8 wk	158.9 ± 8.7 (±SE)
Tail-cuff sphygmomanometer	Male	Wistar	2 wk	∼175 ± 10 (±SD)		Tapia et al. (26)*
			4 wk	∼185 ± 15 (±SD)
Tail-cuff sphygmomanometer	Male	Sprague–Dawley	2 wk	144 ± 3.2 (±SE)		Bidani et al. (66)
			4 wk	134 ± 3.9 (±SE)
Tail-cuff sphygmomanometer	Female	Wistar	9 wk	158 ± 4 (±SE)		Tanaka et al. (67)

SD, standard deviation; SE, standard error; ∼, estimated values.

^*Systolic blood pressure values for this table were estimated from a graphical figure in the referenced study; †5/6 nephrectomy surgery was performed over two surgical steps performed at different times.

People with chronic renal failure exhibit sympathetic overactivity (68, 69), which likely contributes to hypertension in this population (68, 69). There is evidence of increased sympathetic tone in both polectomy (58, 70) and infarction rats (51, 71). For example, Yuhara et al. (70) found that polectomy rats had plasma norepinephrine, a biomarker of sympathetic tone, to be on average 536 pg/mL compared with 230 pg/mL in sham-operated controls 4 wk postsurgery. In these animals, there was a significant, positive, linear correlation between mean arterial pressure (MAP) and plasma norepinephrine (70), consistent with the idea that sympathetic activity contributes to elevations of arterial blood pressure. Similarly, in polectomy rats, Cao et al. (58) found plasma norepinephrine to be, on average, 671 pg/mL in 5/6Nx rats compared with 335 pg/mL in sham-operated controls 10 wk postsurgery. We were unable to find reports of plasma norepinephrine in the infarction rat model; however, Paterno et al. (71) measured elevated renal sympathetic nerve activity in infarction rats relative to normal controls, and Augustyniak et al. (51) found that “neonatal sympathectomy” (neonatal guanethidine injection) reduced arterial blood pressure and proteinuria in infarction rats. The RAAS affects arterial blood pressure and sympathetic tone will be discussed in the renin angiotensin aldosterone system.

Arterial blood pressure is a prominent phenotypic difference between infarction and polectomy rats. Reduced nephron count per se is not responsible for these differences as both models likely have a similarly reduced nephron count following surgery. Although the infarction rat spontaneously develops greater increases in blood pressure than the polectomy rat, blood pressure and blood pressure-related injury can remain a prominent component of pathology in the polectomy rat. Moreover, although the polectomy rat does not spontaneously develop hypertension as the infarction rat does, polectomy rat blood pressure may increase in response to a variety of stimuli, such as a high-salt diet, for example. Investigators may need to carefully consider the arterial blood pressure phenotype when deciding on the use of the infarction or polectomy model to achieve their study aims.

THE RENIN-ANGIOTENSIN-ALDOSTERONE SYSTEM

The RAAS is differentially activated in infarction and polectomy rats. Multiple publications that included prospectively obtained infarction and polectomy 5/6Nx rats have shown significant differences in specific elements of the RAAS (e.g., renin content). These values are shown in Table 3. A study by Ibrahim and Hostetter (15) found that at 2 wk postsurgery, infarction rats, relative to polectomy rats, had significantly greater plasma renin activity, plasma aldosterone, total kidney renin content, SBP, and proteinuria. Ibrahim and Hostetter (15) observed an elevated infarction rat kidney scar (infarct) region renin content relative to the nonscar region. In contrast, polectomy rats had the same renin content at the resection edge as away from it, and both regions had less than the infarction rat scar region (Fig. 2) (15). Others have shown a threefold increase in infarction plasma renin content relative to polectomy rats 3 days following surgery with a correlation between plasma renin content and arterial blood pressure (14). These observations suggest that increases in arterial blood pressure by day 3 in infarction rats are caused in part by RAAS activation. Moreover, this rapid RAAS activation likely contributes to augmented glomerular pathology in infarction rats compared with polectomy rats in the weeks immediately following surgery, as previously reported in these specific studies (6, 14). The findings that similar remnant kidney hypertrophic responses occur between models following surgery despite differential RAAS activation suggests that elements of the RAAS (e.g., ANG II) are not primarily responsible for this hypertrophic remodeling. RAAS activation is associated with elevations in sympathetic activity. For example, increased plasma renin activity is associated with elevations in arterial blood pressure and plasma norepinephrine concentration in a variety of settings (72–75).

Table 3.

Measurements of renin-angiotensin-aldosterone system elements as reported in publications that included direct, prospectively obtained comparison of infarction and polectomy rats subjected to 5/6 nephrectomy

Time Postsurgery	Measurement	Value		Reference
		Infarction	Polectomy
2 wk	Plasma renin content, ng ANG I/mL/h	10.2 ± 8.9 (±SD)	2.5 ± 1.8 (±SD)*	Ibrahim et al. (15)
	Plasma aldosterone, pg/mL	672 ± 591 (±SD)	194 ± 107 (±SD)*
	Total kidney renin content, µg ANG I/h/kidney	127 ± 115 (±SD)	26 ± 18 (±SD)*
	Systolic blood pressure, mmHg	166 ± 31 (±SD)	120 ± 16 (±SD)
3 days	Plasma renin content, ng ANG I/mL/h	59.9 ± 9.6 (±SE)	12.8 ± 1.5 (±SE)*	Griffin et al. (14)
	Systolic blood pressure, mmHg	∼160 ± 10 (±SE)†	∼130 ± 5 (±SE)†
3 wk	Plasma renin content, ng ANG I/mL/h	18.6 ± 3.5 (±SE)	9.3 ± 1.3 (±SE)*
	Systolic blood pressure, mmHg	∼170 ± 10 (±SE)†	∼120 ± 5 (±SE)†

Both studies were performed in Sprague–Dawley rats. SD, standard deviation; SE, standard error.

^*Significantly different than infarction rats from the corresponding time point; †proteinuria values for this table were estimated from a graphical figure in the referenced study.

Figure 2.

Tissue renin content in the infarction rat scar (infarcted poles) regions is highly elevated relative to nonscar tissue and is increased relative to the polectomy rat tissue edge and nonedge renin content. Graphical data were recreated from tabular data (Table 2) from Ibrahim et al. (15).

Interestingly, Griffin et al. (14) observed that by 3 wk postsurgery, polectomy plasma renin dropped to significantly below sham levels (9.3 ± 1.3 vs. 17.9 ± 0.9 ng ANG I/mL/h, average ± SE) and that despite maintenance of hypertension (SBP: 185 mmHg), infarction rat plasma renin content dropped to approximate sham levels (18.6 ± 3.5 ng ANG I/mL/h). This temporal pattern of plasma renin reduction is supported by a Daniels et al. study (50) of infarction rats. Moreover, Ismail et al. (76) observed polectomy rats (two-step surgery with left kidney surgical polectomy performed followed 1 wk later by right kidney removal) to have ∼30% of sham rat plasma ANG II levels 8–10 wk postsurgery. Jackson et al. (77) observed infarction rats to have twofold sham levels of ANG II 1 wk postsurgery but exactly sham levels of ANG II 4 wk postsurgery. Thus, it is possible that elevated RAAS activity contributes to early infarction rat arterial blood pressure increases, but that over time, following volume retention and increased blood pressure, physiological feedback mechanisms enable the maintenance of elevated arterial blood pressure in infarction rats even with ANG II and plasma renin levels reduced relative to their previous peak. Infarction-to-polectomy rat transition experiments in which infarcted poles are later surgically removed may inform upon short- versus long-term contributions of this tissue to RAAS activity. In addition, the presence of the infarcted tissue itself may elicit an immune and/or inflammatory reaction in a manner that polectomy rats would lack due to pole tissue removal. Although markers of renal and plasma inflammation are elevated in both models (24, 53, 58, 78–81), we were unable to find investigations that detailed the specific contribution of the infarcted tissue to an inflammatory phenotype. However, interactions between inflammatory processes and the RAAS are well established. For example, ANG II has known proinflammatory effects such as increasing vascular permeability and expression of proinflammatory cytokines within specific settings (82–84). Thus, the infarcted tissue may exhibit a primary inflammatory reaction as well as a reaction derived from elevated local RAAS activity.

RAAS involvement in infarction rats, polectomy rats, and humans with CKD is demonstrated by the efficacy of RAAS targeting treatments, such as ACE inhibitors and angiotensin receptor blockers (43, 44, 48–50, 52, 55, 57, 85). Part of the mechanism by which RAAS inhibitors function in people with chronic renal failure is through a reduction of sympathetic tone (86). RAAS blockade is a standard treatment for people with CKD and is associated with mitigated progression of proteinuria, increased serum creatinine doubling time, reductions in blood pressure, and delays in reaching end-stage renal disease (43, 44, 87–89).

AUTOREGULATION

The ability of the kidney to maintain a relatively constant GFR, despite changes in systemic blood pressure, is referred to as autoregulation. People with normally autoregulated renal blood flow and essential hypertension are thought to typically exhibit slow developing glomerulosclerosis, because blood pressure elevations elicit reactionary elevations in renal arteriole resistance, thus reducing glomerular pressure and consequent barotrauma (90–92). If blood pressure elevates beyond the autoregulatory range, malignant nephrosclerosis can develop rapidly (92, 93). Parving et al. (94) measured autoregulation in people with nephropathy by quantifying GFR changes in response to acute arterial blood pressure reduction by clonidine treatment. They observed significantly more GFR reduction in people with nephropathy relative to controls, including some individuals that exhibited pressure passivity [ΔMAP (%) = ΔGFR (%)], indicative of highly impaired autoregulation (94–96). Non-nephrotic controls had an ∼2% GFR reduction despite an 18-mmHg MAP reduction, indicating normally autoregulated renal blood flow (96). A recent study by Bidani et al. (56) in healthy conscious rats demonstrated that spontaneous changes in systemic blood pressure are normalized at the renal level within 5–10 s by autoregulation but that normalization is significantly slower in polectomy rats. Thus, even without sustained hypertension, transient blood pressure elevations could cause glomerular barotrauma by slow-acting autoregulation (48). This may be one reason why CKD progresses in normotension, albeit more slowly.

Griffin and Bidani (97) have identified the autoregulation range in healthy rats to be ∼110–210 mmHg and have identified autoregulation impairments in infarction and polectomy rats (6, 9, 56, 65, 98, 99). For example, Griffin et al. (6) demonstrated infarction and polectomy rats to have similarly impaired “static” autoregulation within 2 wk of nephrectomy surgery. This static autoregulation measure is referred to as the autoregulation index and was produced by assessing changes in renal artery blood flow in response to large step changes in renal perfusion pressure, from ∼100 to 140 mmHg (6). An autoregulation index of 0 represents “perfect” autoregulation (no change in renal blood flow despite a renal perfusion pressure change), whereas an autoregulation index of 1 indicates a lack of autoregulation (vessels acting as passive conduits) (6). A lack of difference between infarction and polectomy rats (6) is consistent with the idea that autoregulation is abnormal due to hyperfiltration consequent to reduced nephron count rather than hypertension of the infarction model; the polectomy rats in this study were normotensive and the infarction rats were hypertensive (6). In a different study, Griffin et al. (9) demonstrated that the impaired autoregulation of infarction rats was further impaired by treatment with the Ca²⁺ channel blocker nifedipine, resulting in little glomerular protection, despite reduced blood pressure. In contrast, treatment of infarction rats with the ACE inhibitor enalapril resulted in reduced blood pressure, no change in autoregulation, and significant reductions in glomerular injury (9). The differential effects of these drugs likely occur, in part, by differential autoregulation modulation.

Lin et al. (100) demonstrated impaired autoregulation in infarction and polectomy rats 3 and 4 wk postsurgery. They assessed juxtamedullary glomeruli afferent arteriole diameter at several set perfusion pressures in freshly excised and perfused remnant kidneys (100). Infarction and polectomy rat arterioles exhibited blunted diameter reductions, relative to sham controls, in response to increased perfusion pressure, including virtually no diameter change between perfusion pressures of 100 and 140 mmHg (100). Dilation in response to abluminal acetylcholine application resulted in similar dilation between polectomy and sham rat but a highly blunted effect in infarction rat afferent arterioles, suggesting impairment of endothelium-dependent vasodilation (100). This finding could represent a significant difference between models. That both nephrectomy models had similarly blunted autoregulation despite large differences in the acetylcholine response suggests that acetylcholine-dependent vasodilation has little impact on autoregulation, at least as it is assayed here.

Infarction and polectomy rats spontaneously develop autoregulation impairments. This may be more consequential in infarction rats than in polectomy rats because they spontaneously develop elevated arterial blood pressure relative to polectomy rats, although even transient blood pressure elevations amid normotension could impart renal damage with impaired autoregulation.

NITRIC OXIDE

NO is produced from the oxidation of l-arginine to l-citrulline by NO synthases (NOSs). Although NO has many effects on renal physiology, it is most well known as a vasorelaxation agent. People with chronic renal failure have reduced NO production (101, 102), reduced plasma l-arginine (103–106), and increased plasma concentration of the NOS inhibitor asymmetric dimethylarginine (ADMA) (101, 107, 108). Higher ADMA plasma concentrations are associated with worse disease (101, 108, 109) and are a strong, independent predictor of adverse outcomes in people with end-stage renal disease (109, 110). These observations illustrate the importance of NO handling in renal health maintenance.

Albeit with exceptions, the infarction and polectomy rat models are both associated with decreased NO bioavailability (23, 81, 111, 112). Infarction and polectomy rat studies using urine total nitrite (NO₂) and nitrate (NO₃), stable NO metabolites, as an index of NO production, have reported decreases at a variety of time points (19, 39, 81, 111, 113). Aiello et al. (113) observed a reduction in renal NO production in week 1 postsurgery infarction rats and a near absence 90 days postsurgery, as evidenced by the ex vivo conversion of l-arginine to l-citrulline by renal tissue explants. Ni et al. (23), in 12-wk infarction rats, found significantly reduced kidney tissue NO. Similar to people with CKD, increased plasma ADMA has been reported in polectomy rats (19, 114) and in two-surgical step infarction rats (33). We were unable to find plasma ADMA reported for one-surgical step infarction rats. Moreover, the ADMA concentration has a significant, negative correlation with peritubular (19), and glomerular capillary density (114) in polectomy rats. In infarction rats, renal endothelial NOS (eNOS) levels have been reported as increased (115), not different (111), and decreased (23) relative to sham rats. Similarly, in polectomy rats, renal eNOS levels have been reported as increased (62), not different (116), and decreased (39); the latter two references performed 5/6Nx surgery as a two-step procedure with left kidney polectomy as a first surgery and right kidney removal as a second surgery. Renal neuronal NOS expression has been reported as decreased in infarction rats by Erdely (111) and Szabo et al. (115) and decreased in two-surgical step polectomy rats by Tain et al. (116). Multiple studies have reported elevations in renal eNOS expression in renal biopsy tissue from people with diabetic nephropathy relative to normal controls (117, 118). NO physiology is clearly abnormal in people with CKD. Despite this, establishing a cause-and-effect relationship between CKD pathology and NOS function in people can be difficult. Doubtless, in some cases, intrinsic impairments of NOS function helped precipitate CKD. In fact, certain eNOS polymorphisms are associated with CKD (119, 120). In other cases, the onset of renal disease drives changes in NOS function that contribute to impaired NO production. In infarction rats, Reyes et al. (121) reported that treatment with l-arginine, the oxidation of which NOS produces NO, improves outcomes. Treatment with NOS inhibitors (e.g., N-nitro-l-arginine methyl ester) has been reported to exacerbate pathology in polectomy and infarction rats (22, 33, 111, 121, 122). Studies of l-arginine supplementation in people with CKD have reported little to no apparent effect (123, 124), whereas others have reported modest improvement (125, 126).

Experimental modulation of NO has been used to identify NO-dependent, pathogenic mechanisms endemic to the remnant kidney model, although interstudy methodological differences can complicate interpretation. For instance, Vaziri et al. (39) observed a blunted MAP reduction in 6-wk postsurgery, two-surgical step, polectomy rats relative to sham rats in response to intravenous injection of l-arginine, a blunted pressor response to N-monomethyl-l-arginine (l-MMA; NOS inhibitor, 10 mg/kg iv), and no difference in MAP reduction to sodium nitroprusside, an NO donor. This result suggests impaired NO production in polectomy rats rather than an impaired reaction to adequate NO. In contrast, Griffin et al. (127) reported, in 3–4 wk postpolectomy rats, that l-NMMA administration (50 mg/kg) resulted in nearly identical drops in MAP between polectomy and sham rats. The disparate results of these two studies in the polectomy rat response to l-NMMA treatment, while compatible with each other, could be due to differences in dose (10 vs. 50 mg/kg) or study time points (3–4 vs. 6 wk). They may also be due to differences in baseline arterial blood pressure: despite both being polectomy rats, the rats in Griffin et al. study (127) were normotensive, whereas those in the Vaziri et al. study (39) were mildly hypertensive.

SINGLE-NEPHRON PHYSIOLOGY ASSESSED BY MICROPUNCTURE ANALYSIS

In humans, a GFR below 60 mL/min/1.73 m² for 3 mo is one diagnostic criterion for CKD and below 15 mL/min/1.73 m² is kidney failure. Whole kidney GFR and renal blood flow are initially reduced in infarction and polectomy remnant kidneys (6, 14, 38, 66, 98), but single-nephron GFR and single-nephron plasma flow are elevated in both models 3–4 wk postsurgery (14, 26, 50, 98), indicating compensatory hyperfiltration.

The micropuncture technique enables measurement of individual nephron flows and pressures. Here, we summarize a Griffin et al. study conducted on rats 3–4 wk postsurgery. They reported a single-nephron GFR of 38 ± 18 nL/min in sham rats, 78 ± 9.7 nL/min in polectomy rats, and 73 ± 9.0 nL/min in infarction rats (means ± SE) (14). The glomerular plasma flow was markedly elevated in polectomy (427 ± 39 nL/min) and infarction rats (440 ± 64 nL/min) compared with sham controls (183 ± 18 nL/min) (14). These elevated single-nephron flow rates are consistent with other micropuncture studies (50, 61, 81, 98). Both models exhibited reductions in resistance of glomerular afferent arteries (sham: 1.6 ± 0.2 dyn·s·cm⁻⁵, polectomy: 0.7 ± 0.15 dyn·s·cm⁻⁵, and infarction: 1.0 ± 0.16 dyn·s·cm⁻⁵) and efferent arteries (sham: 1.0 ± 0.1 dyn·s·cm⁻⁵, polectomy: 0.5 ± 0.08 dyn·s·cm⁻⁵, and infarction: 0.6 ± 0.09 dyn·s·cm⁻⁵) (14). Importantly, polectomy rats exhibited elevated glomerular capillary pressures (P_GC; 50 ± 0.9 mmHg) compared with sham rats (46 ± 0.8 mmHg), and infarction rat P_GC (56 ± 1.1 mmHg) was significantly elevated compared with polectomy and sham rats (14). Polectomy and sham rat MAP, at the time of micropuncture, were both 104 mmHg, whereas for infarction rats, MAP was 136 mmHg (14). This finding suggests that not only is arterial blood pressure different between models but perhaps also glomerular pressure. This finding may represent a significant phenotype difference between models. Anderson et al. (50) reported that treatment of infarction rats with the ACE inhibitor enalapril normalized their systemic blood pressure and P_GC with associated decreases in proteinuria and glomerular injury. In contrast, triple drug therapy (reserpine, hydralazine, and hydrochlorothiazide) normalized systemic arterial blood pressure, but not P_GC, resulting in, over time, a similar proteinuria and glomerulosclerosis rate to untreated infarction rats (50). Taken together, these observations suggest that elevated P_GC in infarction rats relative to polectomy rats contributes to their exacerbated glomerular injury as shown in these specific studies. These observations also suggest that the elevated P_GC in infarction relative to polectomy rats is due to transmission of systemic blood pressure to glomeruli and not hyperfiltration per se, as both groups’ single-nephron GFR were similar (14).

Reduced glomerular resistance, as measured by micropuncture in 5/6Nx rats, suggests arteriole dilation. Kimura et al. (128) measured afferent and efferent glomerular arteriole diameter using microvascular casts. In 2 wk postsurgery, two-surgical step, polectomy rats, they reported significantly increased glomerular afferent arteriole diameter (polectomy: 17.9 ± 0.8 µm and sham: 13.8 ± 0.5 µm, P < 0.05) and glomerular efferent arterial diameter (polectomy: 13.7 ± 0.5 µm and sham: 11.5 ± 0.4 µm, P < 0.05) (128). Lin et al. (100) found greatly dilated afferent juxtaglomerular arterioles in both models 3–4 wk postsurgery. In people with hypertension with proteinuria and/or elevated serum creatinine, Hill et al. (129) reported significantly increased glomerular afferent arterioles in renal biopsy samples relative to age-matched, normotensive controls. Like the 5/6Nx rat models, increased afferent arteriole diameter in these individuals was associated with glomerular hypertrophy and sclerosis (129). Collectively, these studies highlight a complex interplay between nephron count, single-nephron GFR, glomerular arteriole diameters, transmission of systemic blood pressure to the glomeruli, and CKD progression.

GLOMERULAR AND PERITUBULAR CAPILLARY RAREFACTION

Capillary loss impairs oxygen delivery, nutrient delivery, and waste removal while promoting tissue ischemia and scar formation (130). Scarring further impairs oxygen delivery, establishing a progressive cycle. Glomerular capillary rarefaction occurs spontaneously in both models (17, 23, 67, 131–133). Interestingly, in polectomy rats, Kang et al. reported an increase in proliferating glomerular endothelial cells, as measured by proliferating cell nuclear antigen (PCNA)-positive cells 2 wk postsurgery. This is followed by a significant reduction by week 8 to below sham levels with a concomitant reduction in glomerular capillary loop number (17). In infarction rats, Floege et al. (134) also found an approximately fivefold increase in glomerular PCNA-positive cells 2 wk postsurgery relative to sham controls, and in 2–10 wk postsurgery, glomerular PCNA gradually reduced until at week 10 it was approximately threefold sham levels. Glomerular capillary loss and glomerulosclerosis are human CKD hallmarks (3, 135, 136). Hohenstein et al. (135), in human glomerulonephritis (of various etiology) biopsy samples, observed elevated glomerular capillary rarefaction relative to normal controls and a linear correlation between this rarefaction and glomerulosclerosis, GFR, and serum creatinine.

Similarly, peritubular capillary loss is a human CKD hallmark and correlates with renal health biomarkers (3, 28, 137, 138). Infarction and polectomy rats exhibit peritubular capillary rarefaction (17, 67, 80, 131, 132). Reminiscent of glomerular capillaries, Kang et al. (17) observed an increase in peritubular capillary proliferating endothelial cells at week 1 but then a rapid decline to below sham levels by week 2 and near absence at week 8. This was accompanied by significant peritubular capillary rarefaction by week 2, with further rarefaction by week 8 (17). Interestingly, the timing of this marked rarefaction increase coincides temporally with increases in proximal tubular damage including dilation, apical blebbing, and brush border loss, as previously described by Chuppa et al. (5) and Adam et al. (139) in polectomy rats. In infarction rats, Advani et al. (131) described loss of peritubular capillaries with the endothelial cell marker JG12 as well as three-dimensional fluorescent microangiography imaging, and Ni et al. (23) described at 12 wk with concomitant tubular dilation and a strong linear correlation with interstitial fibrosis. Similar to human CKD biopsy studies, associations and direct correlations between renal capillary rarefaction and fibrosis have been identified in numerous infarction and polectomy rat studies (17, 23, 131, 133, 140). Glomerular and peritubular capillaries interact with renal tubules in complex ways. Although there is clearly an association between capillary loss with functional loss of tubules, the timing and exact causal relationship between the two and other factors varies with the specific pathological setting. For instance, capillary loss may drive tubular loss and vice versa.

ANGIOGENIC AND ANTIANGIOGENIC FACTORS

Imbalances of proangiogenic and antiangiogenic have been identified in 5/6Nx rats of both models. Although many factors affect endothelial function and growth, we focus here on thrombospondin-1 (TSP-1) and vascular endothelial growth factor (VEGF) because of the extent of their characterization and effect magnitude in progressive CKD. TSP-1, a glycoprotein, facilitates fibroblast activation and inhibits endothelial cell proliferation, among other functions (133, 141, 142). VEGF exerts potent mitogenic activity on endothelial cells and in health is constitutively expressed primarily in the outer medulla, medullary rays, and podocytes (143–145).

Renal VEGF expression has been reported as decreased in infarction and polectomy rats (17, 23, 133). Kang et al. (17) reported significant reductions in polectomy rat tubular VEGF expression by week 4, with additional progressive loss through week 8. These VEGF reductions occurred with reduced peritubular endothelial cell proliferation and increased peritubular capillary rarefaction (17). The antiangiogenic factor TSP-1 was significantly elevated in tubules and glomeruli by week 2 and remained elevated through 8 wk (17). Administration of VEGF to polectomy rats reduced capillary rarefaction, increased the number of proliferating endothelial cells, and improved renal function compared with untreated polectomy controls (132), demonstrating that VEGF can independently alter CKD progression. In a longitudinal study of infarction rats, Hugo et al. (146) observed increases in renal expression of TSP-1 within 1 wk of surgery and steady increases through 10 wk. Collectively, these observations suggest important roles for VEGF and TSP-1 in CKD progression.

Renal expression of VEGF in human CKD is complex and varies with disease and cell type (137, 147–149). For example, Bailey et al. (147) found significantly less VEGF mRNA in glomerular cells in diabetic nephropathy human biopsy samples than controls, but more in people with minimal nephropathy. Namikoshi et al. (148) found that VEGF expression was elevated in biopsy tissue from people with IgA nephropathy with mild/moderate interstitial injury compared with samples with little interstitial injury. However, VEGF levels in advanced interstitial injury were below those with very minimal interstitial injury, despite the highest degree of peritubular capillary rarefaction (148). Analysis of human CKD biopsy samples, of various etiology, consistently demonstrate increased levels of renal TSP-1 (150–152).

SEX DIFFERENCES IN THE 5/6Nx MODEL

Many sex differences in the etiology, progression, and end points of human CKD have been reported (153–164). Women have higher rates of CKD, but men are more likely to progress to end-stage renal disease (156, 165, 166). Despite women having a higher age-adjusted CKD prevalence than men (167), several studies have suggested females are protected from developing kidney disease due to a cardiovascular benefit of gonadal hormones (168–182). However, it has been noted that apparent protection from CKD progression in premenopausal females is lost if the data are adjusted for baseline proteinuria, MAP, and high-density lipoprotein-cholesterol (156). In addition, because of the intimate connection between the cardiovascular system and kidney disease, sex differences in blood pressure and response to chronic hemodynamic stress likely differentially effects renal disease progression between sexes (183, 184).

A unique aspect of 5/6Nx model utilization for exploration of sex differences in CKD is that the initial induction of renal insufficiency is common to both sexes; the renal mass reduction surgery can be replicated between sexes. In contrast, in other models, the nature of the renal insult is itself sex dependent (e.g., hypertension). Thus, use of the 5/6Nx model allows researchers to study sex-specific consequences of renal insufficiency without upstream sources of experimental variation. Despite this, to date, there are relatively few 5/6Nx studies that have included male and female rats. Indeed, the published medical literature as a whole focuses on male subjects disproportionately (185, 186). Within the 5/6Nx female rat literature, most investigations used the polectomy model; we were only able to find four published studies using the infarction model (67, 121, 187, 188). Moreover, very few female 5/6Nx rat studies made assessment of the renal vasculature a point of investigation. For example, we were unable to find any publications that assessed renal autoregulation or capillary rarefaction in female 5/6Nx rats. For the purposes of this review, the paucity of female 5/6Nx publications, especially those with vascular-focused data, greatly hindered our ability to adequately assess sex differences in 5/6Nx rat renovascular pathology. This reality represents an opportunity for investigators to fill these knowledge gaps.

Within published studies that included both male and female 5/6Nx polectomy rats, assessment of basic renal pathology measures such as creatinine clearance and proteinuria in some studies suggested worse pathology in males (18) or worse pathology in females (11, 189). All of the studies referenced in the previous sentence used the two-surgical step polectomy model except for Paterson et al. (11). We were unable to find any published studies that included male and female infarction model rats. Blood urea nitrogen and serum creatinine comparisons across sexes can be difficult to interpret given their dependence upon sex-dependent variables such as muscle mass (190, 191). Clearance measurements (e.g., creatinine clearance) may facilitate comparison across sexes as they normalize for some sex-dependent factors. Some studies used exclusively female rats because they investigated female specific physiology. For example, Antus et al. (168) performed a hormone replacement study in ovariectomized 5/6Nx rats, and Deng and Baylis (187) used female rats to investigate CKD in pregnancy. In contrast, a number of publications used exclusively female 5/6Nx rats (all of these happened to be two-surgical step polectomy models) without providing a clear rationale for the selection of sex (192–195). Similarly, most publications that used exclusively male 5/6Nx rats also lacked an explanation for the selection of sex.

In 5/6Nx polectomy rats assessed 7 wk postsurgery, we found evidence of exacerbated renal pathology in females relative to males (11). For example, creatinine clearance was more than twofold lower in females than in males (11). In addition, through chronic conscious telemetric blood pressure recordings, we found females to have a MAP of 161 ± 7 mmHg (means ± SE) at 7 wk postsurgery compared with 126 ± 5 mmHg in males. Matched sham-operated female and male MAP averaged 114 ± 7 and 102 ± 2 mmHg, respectively (11). Interestingly, despite sex differences in arterial blood pressure, renal fibrosis, as assessed by histological trichrome staining and gene expression, was not different between sexes (11). Furthermore, we found that ovariectomy in 5/6Nx polectomy rats improved creatinine clearance (11). This is likely due to the loss of progesterone rather than the loss of estrogens because it has been shown that estradiol replacement in ovariectomized 5/6Nx rats attenuated indexes of declining renal function, whereas combined treatment of estradiol and progesterone did not (168).

CONCLUSIONS

Infarction and polectomy rats have value in modeling a wide range of CKD types due to shared pathology between these models and people with CKD. Both models spontaneously develop uremia, impaired autoregulation, fibrosis, altered NO physiology, capillary rarefaction, imbalances in pro/antiangiogenic factors, and ultimately renal failure. These two models also exhibit clear phenotypic differences from each other. Infarction rats have augmented RAAS activity. The infarction rat is characterized by rapid-onset, moderate to severe hypertension, whereas the polectomy rat is characterized by slow-developing, mild to moderate hypertension. Differences in blood pressure likely contribute to the apparent faster onset of glomerular injury and proteinuria in infarction rats than polectomy rats. Ca²⁺ channel blockers further impair autoregulation in people with diabetic nephropathy and infarction rats (9, 196). Infarction and polectomy rat models both exhibit increased renal TSP-1, consistent with human CKD of various etiology. Infarction and polectomy rat kidneys consistently have reduced VEGF expression, but VEGF expression is variable in people with CKD. To date, there have been very few 5/6Nx studies that have used female rats for understanding sex differences in CKD. This knowledge gap represents an area of opportunity for discovery. In conclusion, infarction and polectomy rats, despite having much in common, also have major phenotypic differences and thereby model unique aspects of human CKD.

GRANTS

This work was supported by National Institutes of Health Grants R01HL128332 and T32HL134643 and by the Medical College of Wisconsin Cardiovascular Center’s A.O. Smith Fellowship Scholars Program (to R.J.A.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

R.J.A. and A.J.K. conceived and designed research; R.J.A. prepared figures; R.J.A., A.C.W., and A.J.K. drafted manuscript; R.J.A., A.C.W., and A.J.K. edited and revised manuscript; R.J.A., A.C.W., and A.J.K. approved final version of manuscript.