Virulence of Mycobacterium tuberculosis CDC1551 and H37Rv in Rabbits Evaluated by Lurie’s Pulmonary Tubercle Count Method

Abstract

The virulence of the CDC1551 strain of Mycobacterium tuberculosis was compared to that of H37Rv in a rabbit inhalation model. While rabbits that inhaled the two strains produced equal numbers of grossly visible primary tubercles, CDC1551 tubercles were smaller and contained fewer bacilli than H37Rv tubercles. These findings suggest that a miniepidemic near the Kentucky-Tennessee border caused by CDC1551 was due not to increased virulence but to increased transmissibility.

The CDC1551 strain of the tubercle bacillus, Mycobacterium tuberculosis, caused a local miniepidemic of clinical tuberculosis (TB) in the Kentucky-Tennessee border region (25). From 1994 to 1995, in a small rural community with a population at minimal risk for contracting TB, 21 patients were diagnosed with active TB. The source patient and index patient were identified by DNA fingerprint analyses with IS6110 and pTBN12 restriction fragment length polymorphisms. Among 429 TB patient contacts, 72% had positive (often large) tuberculin reactions, including 20% with documented skin test conversions. Only 3% of the noncontacts evaluated were tuberculin positive. These findings imply that the CDC1551 strain is more infectious than other TB strains, and studies with mice (25) also imply that it is more virulent.

Since TB in rabbits closely resembles that found in immunocompetent adult humans (3–7, 16–18, 22), we decided to compare the virulence of the CDC1551 strain with that of the H37Rv strain by Lurie’s tubercle count method (8, 16, 19), which measures the ability of inhaled mycobacteria to establish clinical disease.

Pulmonary alveolar macrophages (AM) of rabbits destroy many inhaled tubercle bacilli before they multiply appreciably (1, 9–11, 16, 20). The AM of humans probably do the same. If the AM fail to destroy the inhaled tubercle bacilli, microscopic lesions are formed. Many of these newly formed microscopic lesions are prevented from reaching grossly visible size by the immune response of the host. (This response also converts the tuberculin reaction.) A reduction in the number of grossly visible primary pulmonary tubercles therefore reflects the power of the host immune response to prevent clinically apparent disease (8).

The virulence of the infecting bacilli can also be assessed by Lurie’s tubercle count method (16). The more virulent the bacillus, the greater its resistance to destruction by both AM and the host’s immune response. Therefore, bacilli of greater virulence produce more grossly visible tubercles.

Thus, with the appropriate controls, the number of visible tubercles present 5 weeks after the inhalation of bacilli provides an accurate measure of both host resistance and bacillary virulence (8, 16, 19, 19a). Also, the sizes of the pulmonary tubercles and the numbers of bacilli culturable from them (Tables 1 and 2) provide additional criteria to assess virulence (as well as acquired host resistance).

TABLE 1.

Characteristics of grossly visible tubercles generated in rabbits that inhaled the H37Rv or CDC1551 strain of virulent human-type tubercle bacilli 5 weeks previously

Rabbit group and no.	No. of inhaled bacillia	No. of tubercles	Avg diam (mm)b	No. of inhaled bacilli required to produce one tubercle	Size of reaction to tuberculin skin test at 5 wk (mm3)
H37Rv infected
01	8,370	4	1.5	2,090	440
02	9,250	4	1.7	2,310	120
03	4,500	2	2.3	2,250	290
04	13,560	28c	2.1	480	220
05	6,160	5	2.2	1,230	220
06	5,710	5	1.5	1,140	270
Mean ± SE	7,930 ± 1,340	8.0 ± 4.0	1.9 ± 0.1	1,580 ± 300	260 ± 40
CDC1551 infected
71	6,930	7	1.4	990	100
72	10,330	0			110
73	5,230	7	1.4	750	70
74	5,160	10	1.3	520	80
75	7,040	2	0.7	3,520	120
76	16,260	5	1.1	3,250	320
Mean ± SE (P value)	8,490 ± 1,730 (0.801)	5.2 ± 1.5 (0.530)	1.2 ± 0.1 (0.007)	1,810 ± 650 (0.743)	130 ± 40 (0.052)

^aThe number inhaled was determined by culturing samples of the aerosol to which each rabbit was exposed for 10 min and measuring, prior to exposure, the volume of air that each rabbit inhaled in 10 min with a whole-body plethysmograph. 

^bThis is the average diameter of all grossly visible tubercles in the lungs of each rabbit. 

^cIt is unusual for one subject to be so different from the others; however, such outliers occasionally occur.  

TABLE 2.

Characteristics of representative pulmonary tubercles produced by rabbits that inhaled strain H37Rv or CDC1551 strain^a

Rabbit group and no.	Tubercle diam (mm)	Tubercle vol (mm3)b	CFU per:
			Whole tubercle	1.0 mm3 of tubercle
H37Rv infected
01	3.0	14.1	300	22
02	1.0	0.5	180	350
03	2.5	8.2	100	12
04	2.5	8.2	7,200	878
05	3.5	22.5	8,150	362
06	3.0	14.1	4,800	340
Mean ± SE	2.6 ± 0.4	11.3 ± 3.3	3,460 ± 1,670	327 ± 142
CDC1551 infected
71	1.8	2.8	132	47
73	1.5	1.8	3	2
74	1.0	0.5	160	320
75	1.5	1.8	2	1
76A	1.5	1.8	1	1
76B	1.8	3.1	6	2
Mean ± SE (P value)	1.5 ± 0.1 (0.02)	2.0 ± 0.4 (0.01)	51 ± 33 (0.05)	62 ± 55 (0.09)

^aData are for 12 representative isolated pulmonary tubercles that were cultured for viable bacilli. 

^bThe volume of the tubercle was calculated with the formula for the volume of a sphere: (4/3)πr³. 

Materials and methods.

The CDC1551 strain (also known as the Oshkosh or CSU93 strain) was obtained from Thomas M. Shinnick, Division of Bacterial Diseases, Centers for Disease Control and Prevention, Atlanta, Ga. H37Rv was obtained from the American Type Culture Collection, Manassas, Va. (catalog no. 27294). Both strains were mouse passaged immediately prior to their use in rabbits. Between receipt and mouse infection, CDC1551 and H37Rv were passaged three and eight times in vitro, respectively.

The number of bacilli in representative pulmonary tubercles (stored at −70°C) was determined by homogenizing each tubercle in a hand-driven Ten Broeck homogenizer and then plating the homogenates in various dilutions on complete Middlebrook 7H10 agar. The agar contained the following antimicrobials to control contamination by other bacteria: trimethoprim (20 μg/ml), polymyxin B (200 U/ml), carbenicillin (50 μg/ml), and cycloheximide (50 μg/ml). These antimicrobials were purchased from Sigma Chemical Co., St. Louis, Mo.

Pathogen-free rabbits (2.5 to 2.7 kg) were purchased from Covance Research Products, Inc., Denver, Pa. At the U.S. Army Medical Research Institute of Infectious Diseases, Ft. Detrick, Frederick, Md., six of these rabbits were infected by aerosols of H37Rv, and six others were similarly infected by aerosols of CDC1551 (3). Briefly, just prior to infection, the tidal volume of air inhaled by each rabbit was measured in a whole-body plethysmograph (3). The number of viable bacilli in an aliquot of the air in the aerosol exposure chamber was measured by drawing the air through an all-glass impinger into an oleic acid-albumin solution (3), culturing this solution in various dilutions on solid Lowenstein-Jensen’s medium, and counting the number of developing bacillary colonies after 3 to 4 weeks. The number of viable bacilli in the air that each rabbit breathed during exposure was then calculated from the tidal volume and the amount of infectious air drawn through the impinger.

After infection, the rabbits were housed in the biosafety level three facilities at George Washington University, Washington, D.C. Five weeks later, the animals were sacrificed, and the number of grossly visible primary tubercles in the lungs was counted as originally described by Lurie et al. (8, 16, 19).

All work with virulent tubercle bacilli was performed in a biosafety level three facility containing a laminar flow hood. All personnel wore HEPA-filtered respiratory devices, surgical gloves, and disposable protective clothing. The two-tailed Student t test was used to determine the P values presented in Tables 1 and 2.

These experiments were approved by the Institutional Animal Care and Use Committees of the three institutions where the work was performed. In conducting this research, the investigators adhered to the Guide for the Care and Use of Laboratory Animals (22a).

Results.

Rabbits that inhaled CDC1551 and H37Rv produced equal numbers of grossly visible primary tubercles (Table 1). The average CDC1551 tubercle was smaller than the average H37Rv tubercle (Table 1) and contained fewer bacilli (Table 2). In contrast to the findings of Valway et al. for humans (25), we found that the CDC1551 strain did not induce enhanced tuberculin sensitivity in rabbits. In fact, it tended to induce less tuberculin sensitivity than H37Rv did (Table 1). In other words, the CDC1551 and H37Rv strains were equally effective in establishing clinically apparent disease in rabbits, and as determined from established lesions, the CDC1551 strain was not more virulent and was possibly less virulent.

Discussion.

It has been known for years that all human-type tubercle bacilli are of lower virulence in rabbits than bovine-type tubercle bacilli (1, 8, 12, 13, 16, 19, 20). Both strains are probably equally virulent for humans (24), but modern human populations are rather resistant to TB compared to primitive human populations never before exposed to this disease (12, 13). Of those people with tuberculin reaction conversions, only about 10% will ever develop clinically apparent disease (2, 14).

In inbred susceptible and resistant strains of rabbits, an inhalation of 70 to 3,000 U of H37Rv was required to produce one grossly visible primary tubercle, depending on the genetic resistance of the host (16). In commercially available rabbits, the number of such units averaged 1,580 (Table 1). The number of inhaled bacillary units required to produce one grossly visible tubercle in humans is not known, but it probably lies somewhere between 20 and 200. Human-type tubercle bacilli sometimes produce progressive TB in humans, but they rarely do so in rabbits (16). Therefore, although rabbits and humans are both quite resistant to M. tuberculosis, humans are slightly less so.

We realize that some M. tuberculosis virulence factors may have more effect on humans than on rabbits. Rabbits are more susceptible to virulent bovine-type tubercle bacilli than virulent human-type bacilli, whereas humans (and mice) seem to be equally susceptible to the two types. Three other laboratories (15, 21, 23) have shown that the CDC1551 strain is not more virulent for mice than the H37Rv strain, and two of these laboratories (21, 23) have shown that it is even less so. Therefore, the lesser ability of CDC1551 to grow in rabbits was probably due to differences in the two strains of mycobacteria and not due to differences in the two animal species.

In mouse lungs, the CDC1551 strain did not grow as well as H37Rv after 14 days because CDC1551 induced a better host immune response (21). In rabbit lungs, smaller tubercles containing fewer bacilli were present in the CDC1551 group after 5 weeks (Tables 1 and 2), which also could be due to a more effective host immune response to the CDC1551 strain. The lower number of bacilli in the CDC1551 rabbits was probably the cause of their reduced tuberculin sensitivity at 5 weeks (Table 1).

The occurrence of the miniepidemic near the Kentucky-Tennessee border caused by the CDC1551 strain suggested that this strain is more infectious than other strains causing TB. Perhaps it has properties that let it survive better in the environment. It is also possible that the source, index, and secondary-source patients aerosolized more bacilli, and/or the airborne droplets that they produced were more stable and of a composition that enabled better bacillary survival. Our study did not address these environmental factors.

In the CDC1551 miniepidemic, the tuberculin converters did not show a greater-than-expected incidence of active TB (2, 25), and the patients diagnosed with clinical disease did not seem to have an unusually severe form. Such manifestations are consistent with our studies in rabbits: compared to H37Rv, the CDC1551 strain had no greater ability either to induce grossly visible pulmonary tubercles or to multiply within such tubercles.